Hi all,
For hot sustained (HS) stress check, most of the times I do OPE - T to save time on checking lifted supports, deleting, rerunning, etc.
Never before I had this situation:
My cold sustained is 80%.
Due to ridiculously high HS stress, 330%, I tried to see what is going on for quite a few hours, and I am still puzzled. I do not have lifted off supports, I do not have guides with gaps. I do have very few springs. When I did 'linearise it' (replace all +Y with Y) and removed spring close to the critical area, calculated HS stress dropped down to 120%.
This does not make sense. Any ideas why this is happening will be appreciated.
Regards
Ranka

Change some supports from +Y to Y for supports you are sure will not uplift. For HOT systems, say > 800 F, load distributions can significantly affect your "Hot SUS" stress. For example, with valves in a riser where the upper support does not lift off but the load in greatly reduced the lower span ends up supporting more weight and thus can overstress the span.

Also remember that the displacements you see for the Hot Sus case are relative to the operating case. Thus you will see negative displacements at some support points. The disp. are NOT absolute, they are relative. I am quite sure this is correct. See for yourself at uplift points. But then again I may be wrong but I doubt it.

Ranka,

There are probably several factors combining to cause the sustained over-stress. The last item of calculated HS dropped down to 120% after spring removed, is probably from the spring hanger being sized with the equipment connection 'free-y'. After removing the spring hanger there are probably large loads at the equipment connection.
The combination of a large weight, a single large spring hanger at some distance from the large weight, an equipment connection requiring minimized loads from piping, and the piping support location having a thermal movement downward, could all combine together in a spring hanger that causes a hot sustained stress that is higher than without a spring hanger. The hot sustained stress becomes worse as the downward deflection causes more force from the spring hanger.
If this is the situation, then I would recommend the spring hanger be located close by the large weight (valve?), re-locate the large weight with its spring hanger farther away from the equipment connection, and consider splitting the spring hanger load by re-designing support to two or three spring hangers.
If there could be three spring hangers, then try to position the spring hangers such that the equipment connection is within the triangle formed by the three spring hangers. This will minimize the moment loads on the equipment connection. If the pipe route is already fixed in location, then this would be difficult.

Thanks very much for your comments.

I was in a hurry yesterday to resolve this, so I did not explain my situation properly.
Today I have calmed down, once again realised that sometimes relatively simple jobs at first sight, can turn nasty, and by rushing, I am not achieving more but less.

I am working on an existing HP superheated steam system. We are replacing an existing 3" control valve in 6" line with a 2" control valve. There is no equipment nearby. 6" line comes of the 10" header, comes down by a nice and flexible routing into a tee rigidly supported. The tee splits into two 3" control valves stations. Then they both go back up via 6" line with rigidly supported bottom elbows, onto the sub-rack. At present there is no springs in this system, it is rigidly supported only.

Now, we are replacing only one of the two control valves. Extra reducers and valve and…

As expected, by reducing cross section I had high exp. stress along short 2" section. So, I extended 2" section into longer length and even had to add small loop to reduce thermal stress. I had high cold sustained, so I added a rigid support close to the top 6" elbow, above 2" loop.

In the model as such, my doubt is if the method of ope-t is giving realistic results!
Run 1. No springs, no equipment, no gaps, only +y supports, none is lifting off HS stress 330%, CS stress 120% (without spring, I added a new spring to reduce CS down to 80%, but this is not important for the comparison I am trying to make))
Run 2. Only by replacing +Y in the model with Y, HS stress drops down to 120% (equal to CS).

My thinking is that this should not happen. +Y or Y should not affect the results in such a huge extent (if the line is not lifting). -Y should not apply the counter force unless the line is lifting! I compered the restrain reports. Support loads (close to the highly stressed area) differ at max 250N, other places much less then that. Where does 120% jump to 330% come from???

For the operating type cases the +Y/Y change will have no effect. BUT for your "thermal only" case numerous supports will uplift since you only have stiffness resisting the expansion, the is no weight included. You are letting a support lift that does not lift. Superposition applies to LINEAR systems. Uplift is non linear.

RS - U have mentioned the support loads differ at 250 N max? is this the difference b/w the normal ope case and sustained case?

If not my reasoning is like this...even though ur supports are not lifting the change in loads b/w cold sustained and ope case causes a redistribution of loads and this causes a redistribution of stresses and in ur case it results in a higher HS stress. When u place a Y support the displacement by which the support moves up (or down) to cause the redistribution of loads is restricted and this causes a reduced stress.

And again the allowable to be considered for hot sustained stress is a topic by itself. I've read somewhere french petrochemical code considers this stress as secondary stress.

Respected forum members kindly correct me of I'm wrong.

I am checking HS stress from the load combination OPE-T, so I was questioning why OPE-T with +Ys is 330% and OPE-T with Ys is 120% when there is no lift offs and since the only difference between OPE modeled with +Ys and OPE modeled with Ys is 250N.

Bob Zimmerman's explanation is correct. If I compare OPE-T with +Ys and OPE-T with Ys restrain report I see much larger difference (then 250N) in vertical loads. Up to 200%

My conclusion from here is that OPE-T is not the best method to check HS stress. Removal of lifted off supports to check HS stress will be better method with more realistic results.

I would like to hear from Coade since they recommend OPE-T method in their Technical Reference Manual.
Ranka

I disagree. Granted most folks are more comfortable REMOVING the uplift supports, but the sustained deflected shape is INCORRECT. The pipe will usually sag at these points, but since there IS A SUPPORT there the pipe cannot sag. The stress magnitude may be comparable but you do not KNOW.

Based on numerous previous posts, people have stong opinions on this GRAY subject, but no one can dispute the shape and thus the strains/stress will be different for the two approaches.

The engineer must provide a sound design that the company and client are in agreement with and satisfied.

Ranka:

I am sorry that I cannot give you an absolutely theoretically proven response on this topic, but simply reveal the way we think at COADE. If you send me (tvanlaan@coade.com) your email address, I can send you a little writeup that I did on this topic, which goes into more detail, but basically here is the condensed version.

In my opinion one real problem is that the sustained and expansion stresses aren't really clearly segregated in an operating case, and depending upon where we place which, the results can be very different, since expansion stress allowables are typically twice those for sustained stresses (i.e., the total stress levels don't change, but how we name them determines the difference between passing and failure). (By the way, SAMAbdul, yes, the French CODETI code does specifically classify the re-distribution of weight stresses due to thermal displacements as an expansion effect).

There are probably an infinite number of possible ways that the system responses (deflected shape) under sustained and expansion loads can combine to equal the total system operating response. To some extent, any "true" distribution of the two types of stress would be arbitrarily determined, but it is our contention that the true sustained stress is most likely enveloped by our "Cold Sustained" and "Hot Sustained" stress results.

It is difficult to say what contributor is due to thermal, and what contributor is due to weight. For example, is a pipe sitting on the restraints due to sagging under weight, and then the thermal displacements pick it up -- or, did the pipe move up due to thermal growth, and then sag back down to the restraints under self-weight? Either assumption provides a different response (deflected shape) under weight, and therefore provides a different weight stress distribution. The illustrative example that I like to give concerns a couple of Emergency Medical Technicians carrying a 250-pound man on a stretcher. The two EMTs are holding the ends of the stretcher roughly three feet above the ground – it is very likely that the stretcher (with the 250 pound man) is sagging in the middle. What is the displaced shape of the stretcher due to weight vs. displacement (lifting it up)? We can model this in these two manners:

1) Cold Sustained – the stretcher was lying on the ground when the patient got onto it. The displaced shape at “installation” (continuously supported, so no real weight stresses) is calculated. Then imposed displacements are applied (lifting the ends of the stretcher 3 feet) are applied, taking us to the operating state.

2) Hot Sustained – the stretcher had imposed displacements by lifting its ends 3 feet high (resulting in a nearly horizontal shape) and then the patient gets on it, causing it to sag from that original displaced position. In this case the Sustained response is the difference from the original displaced position and the final operating (sagged from the original) position.

Now, in order to estimate the distribution of the Expansion vs. Sustained stresses, do I have to know exactly when the patient got on the stretcher? No, because the exact timing is irrelevant – what is important is the effect. In other words, I would be willing to make two bets here: (1) the patient got on the stretcher BEFORE the EMTs lifted it (i.e., case #1) and (2) the actual distribution of the stresses is best represented by case #2! That is the same principle behind COADE’s Hot Sustained solution.

(Oh, and now, let’s complicate things a little further – let’s say that somebody, after seeing that the stretcher is sagging 6 inches under the weight of the patient, shoves a stool underneath that reduces the sag to only 4 inches. Does that make all of the weight stress go away simply because we no longer have a lift-off problem? No, because there still is some sag under weight from the displaced position.)

So let's apply that analogy to your situation with the +Y supports. Some people may want to assume that there is no sag under weight here because the pipe doesn't lift off of the supports. But in reality, the system seems to be growing up thermally, so how do we split the stress distribution -- are we growing up from a well supported weight solution, or are sagging back down from a lift-off situation. If we consider this to be a well supported situation because the pipe has sagged back down to the supports, then let's look at this extreme example:

A pipe heats up and lifts off its support by 1". The analyst removes the +Y support from the analysis, does a sustained analysis, and finds that the stress level fails. Then the analyst discovers that he/she has inadvertantly omitted the fluid and insulation wights (so the pipe actually is much heavier than originally assumed), so it actually sags down further -- until it sits on the support (the stress level would have to be higher in this case, wouldn't it?). Would the appropriate action now be to simply run an sustained case with the support left intact?

COADE's recommended hot sustained solution is meant to provide accurate representations of the weight distributions in all of these circumstances. However, there are still (and probably will always be) questions left unanswered. For example, your specific question here. The hot sustained in the +Y case assumes that the pipe lifts off, except for the sagging effect of the weight. The hot sustained in the Y case assumes that the pipe never lifts up at the support, therefore has nowhere to sag to at the restraint. What's the real answer? I don't know -- probably something enveloped by the cold and hot sustained results.

Hello RS

If you dont mind, can I look at your model. I am really interested. You can send it to this address: fddarbin_mts@intr.mhi.co.jp

Thanks

Normally we don’t check hot sustained stress in a system wherein there are no lift off supports (all supports modeled as +Y and no lift-offs). So there is a possibility that these systems will fail (depending on the temp profile) if we run a hot sustained case. How do we pick up this?