At my company when we model a piping system, we add the structural support stiffness at restraints. Recently in a discussion on which methodology is correct was told that "Industry Practice" was to model all supports as rigid.

I understand that modeling all restraints as rigid is the more conservative approach, but I feel it takes the model a little further from reality. The major problem I have with this approach is that it can lead to overdesigned systems because we are playing it too safe. Finding the support stiffness and entering a conservative value into CII is quite simple so why not.

I would just like some feedback, to see what other opinion are out there…. I know sometimes we get trapped into the “our way is the correct way” mentality. Does anyone know of any good publications regarding different modeling approaches? Image if it is “Industry Practice” something has to be written about this ( I cannot find it).

Thanks

Hi

As a piping stress engineer, I like the sentence "all thing are springs" which is from this forum and the thing is the spring stiffnesses are different. As you said "rigid" which stiffness could be defult, 1e12 in CII or you can define the number you want. In a system, the relative stiffnesses are important.

As you said normally the assumption of rigid supports (stiffness 1e12) is conservative, that means that the rigid point stresses could be conservative, but not all places as the rigid supports take more loads so the other place take less loads in the calculation.

This is only my opnions and I think you can find more discussions regarding this topic in the forum. Most time this point need to be judged based on diffrent systems.

Thanks,
Carter

Hi Martymm

***Finding the support stiffness and entering a conservative value into CII is quite simple so why not.***
If you find support stiffness yes you can use it for advantage & it is realistic also. My opinion is if you want to use support stiffness use it for all the supports in the calculation.
Do you take help from structural guys to find the support stiffness values?
If so it is not much cost effective to get so. More over you need to judge how much practical advantages you are getting for the same rather than using support rigid stiffness ( bit more conservative)
Alternatively you may go for composite analysis( piping + structural model in Caesar).
For some special case finding support stiffness & take it's advantage to pass sensitive equipment nozzle load may be considered.

Regards

Habib

Hi Martymm,
Recently in one of the projects, I used restraint stiffness at almost all the supports in coordination with structural group. I must say that it is not a general practice in my organization to use restraint stiffness at all the locations (also it wasn't in previous organizations I worked with) unless it is absolutely necessary.

As one of the advocates for this on this forum, let me make things clear if I can.

Modeling restraints with the stiffness of support steel is not usually necessary. Pipe racks and building steel is typically many times more rigid than spring cans or hangers, for example. Also, if you find nothing objectionable in your piping system using the assumption of rigid supports, why bother? I (and presumably you, too) get paid to complete projects in as little time as possible consistent with good engineering practice.

When you have pipe with bending stiffness on the order of magnitude of the steel you are attaching to, you may well want to include calculated stiffnesses at your support steel. I have seen pipe racks with 2" of deflection at the center of the span of the member - clearly that is a case where one would want to include the support steel in your model.

I have intentionally left a gray area between cases where it's obvious not to model support stiffness and a case where it's obvious to include it. I cannot and will not force you to accept my engineering judgment - a part of your development as a stress analyst is to understand what portion of this gray area you wish to enter.

Hi Martymm...on most (or all) of our projects we model the fixed supports as rigid. However, we have done some alternate models where we include a structure, or equivalent structural stiffness, and the results can be quite different - sometimes it can result in pump nozzles, etc. being shown as overloaded.

As SHR has mentioned, if you use a specified stiffness for one support, you need to specify stiffness for all support on the system, or your results might be skewed.

Regards,

Pervez

In my opinion this is often due to the steel structure design codes. They often allow a deflection of around 0.004*span which on a large span allows large deflections.
If the support forces are transferred from the piping department to the structural department often the piping experts don't have a look at what structures are engineered to support the piping. My own restrictions for the deflection of piping supports are much more strict than what the codes prescribe just to make sure the pipe stress analysis is still reliable.

BTW Another example: I've once seen a fixed support being calculated with a for steel structures allowable displacement in piping axial direction of around 20 mm.

Hi Corne
Thanks for sharing experience. Can you please make it more clear

*****BTW Another example: I've once seen a fixed support being calculated with a for steel structures allowable displacement in piping axial direction of around 20 mm.****

If I read you correctly you use pipe relative anchor to structure & that structure have horizontal displacement of 20 mm ? If displacement is vertical seems reasonable. Any specific reason of your consideration?
Are you taking something offshore design?

Regards

Habib

Hi Martymm,

just to add another view on this matter:

I have been performing dynamic analyses on pipe systems for 25+ years. When tools and technology advanced over the years it has become more easy to include more details in modelling (such as support stiffness or include complete supporting structures like pipe racks). That is the path of trying to describe reality as accurate as possible.
But there is another path (strategy) as well: use simple modelling (e.g. assume rigid supports) and see what the problems are. When the design is OK then you know that with rigid supporting you can achieve good results. Then recommend to your customer to check (and possibly improve) support stiffness.

This second approach keeps calculations not only simple but also more reliable. You show to the customers how a good (dynamic) design can be achieved and leaves it to the customer how to realize it.

Hello Mr.veen
I registered recently in COADE FORUM,and i was searching the topics of this forum.in your answer you mentioned youself as an stress analyst who doing dynamic analysis for 25 years.honestly, most of stress analysts who i've met have no enough experience in dynamic analysis,and because of that they try to avoid dynamic analysis,but i am so interested to learn more about dynamic analysis and to apply it on some cases which must be considered in dynamic conditions.

a few days ago we recieved a letter from process department that wanted the stress section to consider dynamic effects(vibration) on some lines.
I'll explain the conditions of that line and I hope to recieve your favour to me.

there is a steam header with a branch.a valve is located near to intersection of branch and header.when the valve become open the hot steam comes from header into the branch and because of much temperature differntial between header and branch,there is some condenses at branch.the process department want to be considered the effect of this condense on dynamic behaviour of piping.
I read and search about this but I couldn't find enough answers.I though I should consider acoustic analysis about that.and also existance of condense in steam will produce a condition like to 2 phase fluid.please help me about this problem.

thanks alot.

Mahdikhani

Hello MAHDIKHANI,
I hope the following helps, althoug I realize it is not much:

> I registered recently in COADE FORUM,and i was searching the topics of
> this forum.in your answer you mentioned youself as an stress analyst who
I am not a stress engineer, I am a pulsation engineer.

> doing dynamic analysis for 25 years.honestly, most of stress analysts
> who i've met have no enough experience in dynamic analysis,and because
> of that they try to avoid dynamic analysis,but i am so interested to
> learn more about dynamic analysis and to apply it on some cases which
> must be considered in dynamic conditions.

> a few days ago we recieved a letter from process department that wanted
> the stress section to consider dynamic effects(vibration) on some lines.
>
> I'll explain the conditions of that line and I hope to recieve your
> favour to me.

> there is a steam header with a branch.a valve is located near to
> intersection of branch and header.when the valve become open the hot
> steam comes from header into the branch and because of much temperature
> differntial between header and branch,there is some condenses at
> branch.the process department want to be considered the effect of this
> condense on dynamic behaviour of piping.
I have little experience with that. I can only think of two things: sound velocity will change which should be considered. And, when there is sufficient liquid, 2-phase flow patterns may occur, resulting in a different acoustic behaviour as well.
This really needs attention from a specialist. On the other hand, is there really a problem? Or do they expect a problem? One can perform many types of calculations but they are all waste of time when there is no problem at all.
> I read and search about this but I couldn't find enough answers.I though
> I should consider acoustic analysis about that.and also existance of
> condense in steam will produce a condition like to 2 phase fluid.please
> help me about this problem.

I just reached to this post and wanted to share my thoughts on it.
So, back to the "Support Stiffness" discussion, my suggestion is to treat each case based on the necessities as well as its criticality (experience).

I have recently worked on a piping system and had to consider support beam stiffness where some parallel pumps discharge lines were connected to a header. And the header was resting on beams (sort of piperack). Well, the support beams with the sagging, due to header load, caused higher pump nozzle loads while the pumps were fine with no stiffness taken into account.


On the other hand the pipe support loads should be properly discussed with the structural design engineers to make sense of piping behavior by them as well. The displacement is a good indication and helpful tool that shows piping behavior.

As we know if the pipe axial/lateral loads are simply given to structural engineer to design say a steel T-post you may end up with a perfectly fine sections-baseplate-anchor bolts based on structural codes. But the T-post’s total deflection due to such loads might be more than (still within acceptable limits of structural codes) pipe displacement!
Surely, such deflection of T-post will never happen since the pipe will not move farther than its maximum displacement. In this case, the base/bolts may had to be strengthened due to the given pipe loads while knowing the pipe displacements could have simplified the structural design.

Hence, the mutual understanding of the piping movements/loads by piping and structural engineers will bring some saving to the project.


Regards,

Kamran

If you want to include the support stiffness, then surely the support needs to be designed before you can do this. To design the support you will need support loads. To get initial support loads I assume you would need to define all the supports as rigid anyway......

Unless there was a valid reason for it I doubt most of us have design budgets that allow us to then go back and re enter all the support stiffness values once we have completed a design.

Often we may hand the support design over to another company and have no control over the final design anyway.

In one of the compressor calc, we used a support stiffness of 1.75E+7 N/mm.
In my opninion,
The support stiffness used by C-2 is overly conservative for steel structures(1.75E+11 N/mm)...It could be true for a concrete structure...
For steel the max. stiffness is in the range of 1.75E+7 N/mm and that was the value used in earlier versions of C-2.

So in a calculation using a uniform support stiffness of 1.75E+7 N/mm, if piping is supported on steel structure would result in a conservative and near to realistic support loads..

Using different support stiffness at every location would result in changes in support load distribution....

Regards,

You can change the "default" restraint stiffness in the "Configuration Module".