I tried to model the sample piping system described on Appendix S and my observations are as follows (with +Y removed):

1. lift off at node 50 is about +18 mm (case 1: W+T1+P1)
2. sag at node 50 is about -53mm (case 3: W+P1)
3. sus stress case failed (equal to the so called “hot-sustained")
4. ope stress case passed (per Appendix P)

Question:

Can we say that the sample piping system stresses are still in compliance with the code given the following arguments?

1. CAESAR II calculates firstly the displacements, then the forces & moments, lastly the stresses. Stresses are based on displacements, hence sustained stresses are based on sustained displacements, and i.e. sustained stress failure is due to -53 mm sag.
2. In reality, the -53 mm sag is impossible to happen because there is a physical support that exists below the pipe that lifted off; hence the sustained stress calculated is also impossible to happen. The proof is if node 50 is allowed to sag by -18 mm only, the sustained stresses are below the allowable. Note that the amount of lift off is just +18 mm above the physical support.
3. In reality, after the pipe lifted off by about +18 mm, there will be a redistribution of loads, the resulting operating stress will be calculated based on the +18 mm lift off. To comply with the code this operating stress should be lower than SOA and all other operating stress and operating stress ranges should be lower than SOA (per Appendix P).


Thanks,

Looks like you're talking about Example 2 in App S, B31.3.
Regarding your point 1, sustained stress is based on weight, pressure and sustained forces; displacement is not normally considered.
Re point 2, I've heard that argument before but have to disagree. The reference point for hot sustained has to be the expanded position of the weightless pipe (ie a case with T1 only). Based on your figures that displacement at node 50 would be about 18+53=+71mm. When weight is added the pipe sags down to +18mm.

Thanks Mr. Martin but regarding argument no.1, I am talking about the chronological order of calculation used by CAESAR.
CAESAR calculates first the displacements followed by the moments & forces followed by the stresses.
The so called "hot-sustained stress (W+P1)" is based on the -53mm sag from the original position, that's why the piping failed.

I don't have the Example 2 model in front of me to confirm your numbers but I can respond using your data...

If the pipe grows up 18 mm in the operating state and the pipe sags 53 mm under weight alone with the restraint removed, the pipe would grow up 71 mm in a weightless environment. That 71 mm is caused by the thermal strain. When you add in the deadweight, the node now lifts up only 18 mm.

1. I'm not sure where you are going with this displacements give forces & moments and forces & moments give stresses. That works with any CAESAR II analysis. You find the global position of each point in the system then back calculate the local forces and moments and, from those, the stresses. Do not confuse these general displacements with "displacement loads" or "strain loads" or "expansion loads". Displacement loads are evaluated based on a fatigue limit, stresses due to sustained loads are related to system collapse.

2. You are right, the system cannot sag with a +Y support in contact with the pipe but when the pipe heats up, the support disengages from the pipe - by 71 mm (using your numbers). There's plenty of room in the hot position to allow this sag.

3. Do not confuse Appendix P's operating stress evaluation with the other B31.3 requirements for stress. Appendix P is Alternative Rules for Evaluating Stress Range. It has nothing to say about satifying the longitudinal stresses due to sustained loads.

Answer:

If you agree that the liftoff at node 50 is a deadweight effect, then, no, the system is not compliant with B31.3. See S302.8. I would claim that the liftoff at 50 is a thermal effect. I would claim that if the system tends to collapse , that the support would once again become active and once again adjust the logitudinal stresses due to sustained loads. So, I would claim that the system will not collapse and is therfore compliant with the requirements for longitudinal stress due to sustained loads. An issue that I think several engineers have is the magnitude of the liftoff. A little (undefined) is OK (perhaps marginal design) but a lot (undefined) is poor design.

right or wrong,
Design a system where pipe shoes lift off the supporting structure and you will get a phone call from the plant manager!
Or somebody will fill the gap.

The secondary effects of engineering design!!!!!!!

Thank you Mr Dave,

Please forgive me, I just want to make this follow-up.

On the following excerpt from B31.3 Appendx P, why was "pipes lifting off..." particularly mentioned ?

"P300 GENERAL
(a) This Appendix provides alternative rules for evaluating
the stress range in piping systems. It considers
stresses at operating conditions, including both displacement
and sustained loads, rather than displacement
stress range only. The method is more
comprehensive than that provided in Chapter II and is
more suitable for computer analysis of piping systems,
including nonlinear effects such as pipes lifting off of
supports."

Also, I attached something to clarify my views. Please comment.

Thanks,

Julius,

This set of alternative rules concerns displacement stresses combined with other loads. The requirements for sustained load stresses in the base code still must be met.

So first and foremost this distinction must be made.

Now why and how of supports that lift off.

First I have an article discussing Sustained stresses in one of the COADE newsletters I suggest you read that for background.

The committee (B31.3) has repeatedly said that sustained stresses must be evaluated in all various support configurations, in other words lifted off, partially lifted off, and in full bearing.

The "HOW" we have never defined (nor will we never I suspect) because of the complex nature of the issue.

My own opinion differs from COADES on this subject I would run the analysis sans the support and see what the sustained stresses are. Also for high temperature or extremely cold systems I would use the Modulus E at temperature to examine the systems operating cases to make sure that the liftoff does indeed occur Mod E may vary significantly!

I hope this helps its a tough subject but when one considers how easy it usually is to avoid these problems we should try to do simple things first ... move the damned support, use a spring or simply get rid of it if possible.

Julius,

Why was "pipes lifting off" explicitly mentioned in Appendix P?

There are two stress limits provided for the expansion stress range - (1a) & (1b). (1b) is the most complete and includes the value SL - the longitudinal stress due to sustained loads. (1a) is simpler and conservative by setting SL to it's maximum allowed value Sh. If I wish to use the higher allowable limit for the expansion stress range, I have to know what SL is. With nonlinear supports, longitudinal stress due to sustained loads can change so it may be difficult to assign a specific SL in setting the limit for each of the several expansion stress ranges. Appendix P removes that dependence on SL in setting the allowable stress range and allowable operating stress.

So, systems with +Y supports may have liftoff and liftoff might be more easily evaluated if SL is not an issue.

On your illustration...

In your first image (without the support) you show a +Y displacement in the operating position and a -Y "sustained" position. What you do not show is the position of the line in a "thermal only" position - as if you take the system into a weightless environment and heat it up. It is this displacement that equals about +71 mm. (I'm not looking at your figure now,if I recall you show +18 mm in operating and a sag of -53 mm. This is where I get the 71.)

Our focus has always been on the expansion range. Our L1-L2 approach does that well. On the sustained side, like John suggests, you may want to examine the sustained loads without that support. It looks like your thermal growth upward exceeds the pipe sag downwards so that the support will disengage. You can run a second W+P analysis without the support or you might run a "thermal only" (T1) case and subtract that from the operating case to back out the longiudinal stress due to sustained loads. John wrote a newsletter article about this and I put together a short ASME PVP paper about it two years ago.

Thanks Mr. Luf, I already read your article a long time ago and in doing my analysis, I always make it a point to remove the supports which lifted off for a check of the "hot sustained stress". Also, if possible, I also try to avoid such cases by relocating supports or by adding flexibility to make the support sit.

Thanks Mr. Diehl, it is clear to me now that Appendix P is written for the purpose of checking NOT the sustained stress but the displacement stress range by means of checking the Operating stress and the Operating stress range. In addition to what you have explained regarding the reason behind the explicit mentioning of "lift-off", you will notice that the operating stress is the same for both model 1 (+Y removed) and model 2(+Y retained).
Hence, the displacement stress range can be checked in either model just by checking the operating stress.

I did not place the T1 case because I thought it was not realistic, i.e. being weightless. I prefer the L1-L2 case, i.e. the expanding mass of piping. By looking at model 2, I thought collapse will be hindered because
of the actual existing support that will limit the sag, preventing the pipe from assuming the failure-shape shown in model 1. This is why I made model 3 which is a "what-if" case to check the sustained stress if the pipe sags by only 18mm and then sits. But I guess I was wrong. I will just continue with my previous ways of removing the support which lifted off of supports.

Thanks,

Hello Mr. Julius,

This is a very interesting discussion…

I have read discussions on hot sustained, Luf sirs article and also Coade’s writeup on this subject…

Somehow I could not agree on Coade’s point of “thermal growth of a ‘weightless’, non-pressurised pipe…”, since this is never a reality. Their argument is that, exact timing of the loading is insignificant but only the final effect. Hence the +71mm Dave mentions will never occur in reality!

So as long as we check against the basic Sh allowable limit, CAESAR II’s hot sustained analogy may not be the perfect solution. On the other hand if we agree that hot sustained has secondary stress properties and check against a higher allowable (as CODET1 suggests) CAESAR II’s hot sustained has more significance.

It again leaves with me Julius’s question…why can’t we do a sustained stress check by not removing the support but by introducing a gap of 18 mm below?

See my response in the attached zipped file

Thank you Luf Sir for your detailed reply.

But once we observed some erratic results (in terms of diaplacement in hot sustained case) in a big model with +Y supports...so we always prefer to do a check by removing the lift off supports.

Another interesting case was in another model, all +Y supports, with NO lift off in any of the supports, showed hot sustained failure!! In my opinion this is because hot sustained stress check picks up load re-distribution effect because of temperature and so is the failure...So many of our on-lifting supports model might fail because of the load redistribution effect!!

But Sam, that's the thing - it's not simply lift off or no lift off, it's how much the thermal effects shift the load elsewhere. The longitudinal stress due to sustained loads does not "jump" from a value calculated from the weight analysis with the support active to a second value from the weight analysis with the support inactive, it's a gradual shift just as the support load (usually) drops as thermal strain is added to the system.

Yes Dave, everyone starts watching only when there is a lift off...even if the load changes 99% (bcos of temp effects), unless it lifts off no one bothers...

So do we need hot sustained check for all models??

The Code is not clear on this point. The latest B31.3 interpretation on the matter states that longitudinal stresses due to sustained loads should be examined for all support conditions. As you point out, we are alerted to a change in support configuration when lift occurs. But these stresses change as the system's thermal strain changes, lift off or not. The Code, using "slide rule" methods, does not address this issue. With tools like CAESAR II, we can generate better numbers, but the Code provides no "guidance" on how to use them.

My life is simpler if I assume any change in load or stress due to thermal strain is considered in the expansion stress range. That's what we have here. Supports for deadweight are usually handled by rule rather than by analysis - that leads me to believe that we're putting too much work into this subject. It's an interesting subject though...

I have recently experienced a system where there was a hot sustained failure, however the OPE-T displacement was horizontal, there was no sag.

The failure occurred at a flat turn (30 deg bend-the offsites rack is completely skewed) in a pipe track. It is an existing rack and where possible (given real estate constraints) we have used vertical legs to achieve non-45/90 horizontal changes in directions.

Looking at the displaced shape (at the failed node) for OPE and T, there was a large difference in the lateral and axial positions, but not vertical. So the stresses caused by that displacement were measured by the sustained allowables.

The normal expansion and cold sustained stresses were well within the code allowables.

Should this stress be measured by the hot sustained case?

"I have recently experienced a system where there was a hot sustained failure" What was the nature of said failure.... crack, collapse? Post a photo please!

If your measured displacements at the failure location in the operating condition were significantly different than those in your analysis, it indicates one of two things.

1. Something was missed in analysis (in which case the analysis results as a whole are less than worthless).

2. Something distant from the observed failure (such as a shoe, guide, anchor, or axial stop) has failed so that the analysis, although valid for the design conditions, has ceased to become so.

You have a lot of field work to do. Enjoy!

To clarify, I am talking about the results of a stress analysis, not the actual failure of a physical piping system. This system has not been built yet.

If a hot sustained case was not performed, there would be no indication of any overstress as the cold sustained and expansion stresses are well within the allowables. There is no lift off at any of the nearby supports.

The point I am trying to illustrate here is that I am seeing a hot sustained failure that appears to be due to the horizontal displacement of a bend. There is no lift off in the T, or OPE case. Cold sustained, and expansion stresses are well within allowables.

The hot sustained case is a "construct" therefore its displacements are not to be taken as something that could ever be seen a in real life.

Hi Dude,

Is it possible for you to attach your CII files here? I am interested to see your system, although I dont know why you checked the OPE-T case, when there are no lift-offs.

Thanks,