Pressure vessel nozzle loads 3

[Aberswn1982]

My question is regarding assessment of defects in pressure vessels and defining accurate loading boundary conditions; If I have to assess (be it hand calculations or FE analysis) a vessel with an identified defect (say internal corrosion - local thinning), how can I obtain suitable nozzle loadings from connecting pipework?

I guess the correct approach would be to model all pipework and supports systems (under the relevant load cases) in a stress analysis and determine the reaction forces/moments at the nozzles on the vessel? However i feel the time spent on modelling the pipework under test pressure, thermal expansion wind load etc is all in vain as more often or not the pipe work will never match up on site and always has to be force into position to mate up the flanges i.e. we are inducing fit-up or long-range residual stresses, which would ultimately affect the outcome of the pipe work stress analysis. Is there a rule of thumb or general approach for applying boundary conditions onto nozzles i.e. for a vessel analysis just apply the max loadings for that flange class/size as given in the parent code?
any thoughts?

 

[TGS4]

for a vessel analysis just apply the max loadings for that flange class/size as given in the parent code

Just out of curiosity - what specifically do you have in mind in this regard?

You are correct that fit-up loads would not be accounted for in a standard pipe stress analysis, but usually we're dealing with either large self-weight (gravity) loads or large thermal (restrained free thermal expansion) loads.

I wouldn't bother with the test condition, unless the attached piping will actually be tested...

 

[Aberswn1982]

For example, internal corrosion found on the bore of a nozzle on a horizontal, saddle supported vessel. Nozzle is a 24inch opening and connecting pipework is a long pipe run taking the contents from the vessel off to another part of the plant.

The corroded nozzle remaining wall thickness may easily satisfy the minimum required wall thickness for pressure containment, but how can i analyse the stress in the nozzle bore wall and obtain an estimate required wall thickness for structural integrity - i.e ability to carry weight of connecting pipework. I could dimension the piping system but supports will rarely be all connecting/seating correctly and hence inaccurate - how do i account for this in applying nozzle forces/moments in the stress analysis of the vessel nozzle?

this is an example I have made up but scenario can be common. Just interested in how other people define boundary conditions to apply in an analysis of a vessel nozzle. I see from design/new build, piping systems are modelled and forces/moments obtain and used, but for in-service 'old' equipment what is a suitable approach?
thanks
Ed

 

[gr2vessels]

I would check first the original design calculations, to see what loads have been included in the original design. It's probably no point to re-evaluate the nozzle for unrealistic nozzle loads, which were not included first place. Also, the design calculations might reveal what nozzle loads have been included for the vessel in corroded condition ( I assume the defect/corrosion mentioned is within the corrosion allowance). If the corrosion exceeds the allowance, then perhaps a repair and a API579 analysis would be required before any nozzle capacity evaluation.
Cheers,
gr2vessels

 

[Aberswn1982]

hi gr2vessels,

thanks for the reply, design calcs? The plant is 40 years and documentation is poor. In some cases, definitely not all(!), we may have a GA drawing if we are lucky, definitley do not have the design calcs. Plant ownership has been numerous and handed over many times, orignal design calcs we just do not have, hence the post, where do I start to obtain conservative, maybe not accurate loads and a useful/meaningful analysis.

Also, this would be where orignal design corrosion allowance has been consumed but there is still inherent safety factors and hence some more corrosion allowance available until tmin is reached, so definitely donot want to jump straight to repair.
thanks
Ed

 

[racookpe1978]

D*mn. Tough problem.

You've got a specific problem area (problem nozzle or nozzles) in a particular place, but at least it has not failed. Yet.

Apparently you have got a budget, or a job order or a directive from the boss to look at the problem, but that's not an infinite source of infinite analysis of infinite possibility of different problems. And, regardless, only one of the problem analysis cases is actually correct. The rest? Either wrong or or irrelevant and a waste of time.

Correctly, the theoretical nozzle stress analysis assumes pipe load to the vessel is zero-zero after the pipe hangers are correctly hooked up and pre-stressed. Zero load, zero moment that is, at normal pressure & temp.

So. To save time, disconnect the biggest nozzle-pipe connections. Look at the moment (movement actually) at ambient temps and zero pressure. Massive movement means massive stress when the pipe is re-hooked up ------ IF the movement from NOT/NOP is taken into account. You will probably have 2 or 4 big nozzles + big pipes. Check those for "real world" stress applied back to the vessel at NOP/NOT.

Fix the hangers and pipes so the biggest nozzles have as close to zero-zero stress-moments at NOP/NOT as possible.
For the corroded nozzles, adjust/rebuild/rehang the pipes where ANY corrosion and problems are expected. (Remember, this is while you are trying to get zero-zero stress at NOP/NOT not at room temperature and atmospheric pressure.

So, you've released the biggest pipe stresses from the nozzle + pipes. Look very very carefully at the 1 or 2 nozzles that are corroded, and do the FEA checks on those 1 or 2 nozzles using the internal pressure/temp stresses only. Your problem then is reduced to known "holes" (hollow spots) in just a few places.

 

[TGS4]

When I do this type of evaluation (granted, I'm usually called in to the Level 3 API579 assessment), I simulate the piping using the best available information on the supports. Following racookepe1978's advice, otherwise, is VERY good advice.

If you are performing an API579 assessment, you should be getting more accurate the higher your Level. I would make conservative assumptions for a Level 2 (although if memory serves me correct, by the time you add nozzle loads to a LTA, you're automatically into a level 3...).