Pipes, not me.!

I have a system which has had two trunnions (on a letdown station) anchored down as opposed to a stop and guide as requested on the calculation. this system is currently in operation (hot)

Caesar gives overstress of 300%(610n/mm2)
Kelloggs hand calcs suggest stress levels of 3600n/mm2 on the trunnion.

(JCL, Anchor nodes 20400 & 20500 on the CAESAR II file i sent)

I am thinking these stress levels are indicative only, and that local yielding, anchor slippage (base plates are bolted) and some minor buckling will reduce the actual stress levels somewhat.
Ideally, i would like to use FEA to asses what is going on, but i dont have it so my question is, at what (over)stress levels do i raise the alarm?. I have requested that the anchors are removed and the spools local are replaced with immediate effect for safety, but at what point/stress level can you expect rupture or catastrophic failure given all of the factors of safety invloved?
The Min UTS as per App.A for API5L is 410n/mm2
The trunnions are way over this level (as is the Caesar file)

Or. can i actually give a technical answer due to the vagaries of local yielding and deflections.

answers on a postcard..........

So for those PSI people out there this "3600n/mm2" is an extremely high 522,144PSI (Thanks to Rich Ays unit conversion program a free download on this site)

Although as you point out local yielding can and will occur there must be limits on this type of affect. Section VIII Div 2 gives stress limits based upon the type of load as well as making sure that primary or sustained loads are adressed before adressing secondary fatigue type loads. However for fatigue loads you can only go up 3Sm and your Kellogg number is far above that. (Take a read on Section VIII DIV 2 for a more specific information)

"....at what (over)stress levels do i raise the alarm?." UGGGHHHH the problem is you have no way of knowing what the real localized stresses are. If for instance I had complete confidence in Kellogg I would be flag waving now.

"but at what point/stress level can you expect rupture or catastrophic failure given all of the factors of safety invloved?" FEA as well as in situ review of what was actually built may give insight on this extremely difficult question but without FEA I do not believe this can be answered, only an educated and speculative guess can be made.

Sorry for the vagueness...

Superpiper,
I am with "in situ' check.

A hard one Superpiper.
I'm guessing the high loads you are seeing are secondary in nature rather than primary? What happens to the load if you tweak the anchor stiffness, gaps etc to something perhaps more representative of what is installed. I'm also with Anindya on getting a field check done. Unless it is a very short trunnion, any overload will probably be immediately apparent from the trunnion being 'on the p*ss'.
If you want to do some digging, there may be a usefull clue in WRC 448 however. They did a series of limit load tests based on when the measured deflection reached a deflection of twice that predicted by elastic behaviour. Perhaps not exactly what you have but never the less a pretty substantial loading I would have thought. They then report that the load limit can be described as the [yield stress x section modulus] / [2/3 Cx]; Cx as described in N-318-5.
It might be worth looking closer at this to establish if your trunnion is loaded beyond a comparable limit to this remembering this load test was done with a primary type load [even if some the local effects can be regarded as secondary in nature]
Best of luck anyway.

"I am thinking these stress levels are indicative only, and that local yielding, anchor slippage (base plates are bolted) and some minor buckling will reduce the actual stress levels somewhat."

Best practice is never to design "into" the factor of safety: Your calc's should keep the predicted stress levels below the highest stress you want the metal to see. So the actual stress "should be" less than the yield stress/factor of safety.

But, if the actual stress anywhere is too high, and it is too high IF your calc's are right, then the real metal yielded. Somewhere, we just don't know yet. The yielding will continue at some point until everywhere the max actual stress is less than the yield point. (This is what finite element analysis simulates actually, in this case it seems to be happening in real life. Probably a combination: bolt slippage, plate bending, trunnion "local squishing", trunnion mount-to-pipe bending, pipe bending, etc. Each failure (yielding) reduces the stress on the whole assembly but increases it somewhere else on someother component until that piece yields and shares the load elsewhere.

On the other hand, any chance that the initial load, or the initial conditions of movement, aren't as much as you predicted?

"The yielding will continue at some point until everywhere the max actual stress is less than the yield point."


Or if things are too extreme the yielding continues until there is a failure! Thats why staying in the code atress level is a lot easier on the mind!


I always say not to worry about anything being overloaded because if its really terrible once it breaks the overload goes away!

The point of this thread was:

1.By theory my pipework is overstressed
2.It hasn't broke yet
3.It may or may not break ever.
4.What correlation is there betwix reality (site) and fantasy (my model)


Going forward, i'm not bothered what damage has or has not been caused, only that the potential is definatly their. So on this point i have instructed site to correct the supports and replace the spools and trunnions local to the overstress.

I remember my ex-boss blowing smoke up the engineering managers ar$e by saying that stress levels on a model were above the failure stress.
at the time i thought it a daft comment, and i am now reminded of it.
In this situation, the levels of stress calculated are above the UTS for the material, however, failure has not ocoured (due to yielding/self limiting etc,etc)
I would not try to accept or design for these stresses either, any justification(s) possible due to yielding/slippage/etc etc is outside my sphere of confidence and i ain't going there.

But, These overstresses have not caused failure of the pipe, but they are causing a local ball valve to "pass" fluid in the closed possition. i am assuming the local loads are exccesive.
This leakage IS going to make me change how i review stress levels. Because Caesar is showing no overstress at this valve, i am going to be more cautious of high forces on inlines items. I previously assumed that a valve was perfectly capable of handling any stresses given it bu the pipe
(within reason)

edit(mi speling is atroshius!)

PSV manufacturers actually publish load limits on their valves... also some brewery valve manufacturers do so.

But overall load data is hard come by...


Seek and maybe you will find something.... don't seek and I guarantee you will find nothing!

Good Luck in your Endeavour

SUPERPIPER - I hear this all the time from clients, "if it hasn't broke how can it be overstressed?" It is very difficult to predict failures, as you know the Codes are very conservative. Depending on the failure mode, by a factor of 4 to 8.

If you are looking at secondary loads, the SIF on a trunnion can be as high as 10-14, which is a "cycles to failure factor" not a "stress concentration factor". Note SIF's are not really accurate for trunnions, but it seem's to be common practice. I generally check the trunnion loads seperately with Kellogs or Roarks and don't put an SIF in the model. I also pre-calculate maximum trunnion loads and put them on the support standard detail.

So if Caesar tells you the stress is 300% that doesn't mean the stress is actually that value...it's just a Code compliance ratio, not really a stress ratio. FEA would more accurately calculate the real stress ratio. On bends I've done some FEA and the stresses were much lower than "predicted" by the SIF.

If it's a primary type load, failure would occur when enough of the cross section is beyond yield. There's a good example of this in the user guide on the FE/Pipe website that show's a cantilever beam loaded with a secondary load and produces a plastic hinge, and second model with an equivalent primary load that produces yield through the whole cross-section and predicts collapse. I think the FE/Pipe user guide is downloadable by anyone, or it used to be.

Engineers worry when their mathematical models do not coincide with reality. Physicists worry when reality does not coincide with their mathematical models. Mathematicians worry when they see others trying to apply their mathematical models to reality. From this I deduce that Superpiper is an engineer!

The leaking ball valve appears to me to be a big clue. Nature always finds the weakest link to attack. I once had the privilege of listening to a great structural engineer explain to a client that there was no point in continuing to replace rivets that had already popped out several times, because, "Mother Nature is telling you she doesn't want you to have rivets there." (Younger engineers reading this should perhaps consider the fact that he already had done plenty of analysis to prove that the overall strength of the joint, minus the absent rivets, still met the design code.)

The trunnion, although overloaded in theory, is likely to be able to bear considerably more load before failure than the ball valve body. And the leaking ball valve is likely being distorted by the loads that CAESAR II thinks are being applied to the trunnion, so that the calculated loads on the trunnion are not a part of reality.

It might be interesting to model the ball valve as flange pair with a short length of pipe with OD and wall comparable to the data presented in table 3 of ASME B16.34 between the flanges. You might find this to be a whole lot more useful model for a valve than simply calling it "rigid," at least for the problem at hand.

As to the recommendation of a site visit, I heartily agree, but I would try to be very, very careful and alert near the trunnion. Most of the time, I would expect the trunnion to simply shear off the pipe. But most of the time drunk drivers manage to get home safely. Our design rules are based to some degree on statistical data, and statistics are only averages massaged by mathematics.

Obviously, as noted by many above, the load type is absolutely critical to any analysis of this problem. Your first reaction seems to be very prudent - get the load off the offending structural element first, then figure out how to replace that structural element with a non-offending one.