Piping on API Pump skids

[Janimi]

Hi guys,
I am wondering about the design of most suppliers for API Pump skids especially API 681.
In most cases it is required to keep pipe length as short as possible for a good process. When you hit the google search you can find a lot of examples for those pump skids with Vessels, HX, etc.
Do those suppliers not have a detailed look on the piping design regarding thermal expansion? In most cases acceptable nozzle loads are very low and stiff piping will result in high forces an moments. Pumps and vessels are directly bolted without any sliding saddle. Expansion joints are often forbidden in oil and gas market and from my point of view also rarely used because of limited life time.
On the other side the pipes are very short and temperature not high so there is not so much expansion.
How do they handle this practically? Just do not care about pipe stress calculations or make some special assumptions and ignore requirements of e.g. B31.3?

Let me know your feedback

Thanks

 

[XL83NL]

Though your points or concerns seem to make sense, what doesnt make sense (to me) is that API 681 (version Im looking at is the reaffirmed 2010 edt) points out, in para 3.7, that piping needs to conform to ANSI B31.3. Though ANSI has been replaced by ASME B31.3, following that standard should result in a compliant design, from thermal displacement stress point of view, as well as nozzle loads for connecting equipment.

You could always ask a supplier ifthey can show compliance to B31.3 for cases where you have your doubts. I wouldnt be suprised if the piping design by such suppliers is limited to a (simple) pipe spec only, covering only the wall thickness calculation at its best.

/edt; pls dont double post. As this thread has had the 1st reply, Ill RF the other thread for deletion.

 

[Janimi]

You're correct but looks like a typo to me, because in para 2.9.3.3 refers to ASME B31.3 and not ANSI. (The references para 1.5 also shows ASME, not ANSI)
Nevertheless it is required as per para 2.1.12 to keep piping loads into account even if you do not follow any kind of piping spec...

*edit: other post is deleted.

 

[XL83NL]

Do you maybe have an example of an (exisiting) skid hwere you have doubts regarding compliance to API 681, thus compliance to B31.3?

 

[Janimi]

Sure, please find attached some examples from google..
Have a look on the discharge pipes between pump and vessel.

 

[XL83NL]

I see what you mean. Perhaps the manufacturers use B31.3 para 319.4.1(a), and possibly (b) to 'judge' their system satisfies B31.3 requirements for flexibility. Although I would argue that most manufacturers dont see the result of 20 yrs lifetime for the piping systems they design, hence cant (always) use that section of B31.3 for jusitification.

 

[RVAmeche]

I've always wondered about this too. We buy a piece of equipment (tank, pump, etc) and the vendor almost always says "no loads allowable!" and refuses to give a number.

Then you can buy a skid like the ones above and everything is okay? I assume the skid supplier coordinates with the equipment vendors to ensure the nozzle loads are acceptable but that may be optimistic.

 

[KevinNZ]

API allowable pump nozzle loads are very low. Hard job to satisfy thermal loads (flexible piping) and seismic loads (restrained piping)at the same time.

Even harder when the allowable loads do not increased with the flange class. A class 900 pump body should be stronger/stiffer than at a class 150 body but the allowable loads are the same.

 

[XL83NL]

We quite often see pumps where the MFR has a nozzle load table on the GA drawing, in which they mention all loads and moments shall not be greater than 0 N resp 0 Nm.

 

[Janimi]

I see you are struggeling with the same problems like I do

I see what you mean. Perhaps the manufacturers use B31.3 para 319.4.1(a), and possibly (b) to 'judge' their system satisfies B31.3 requirements for flexibility. Although I would argue that most manufacturers dont see the result of 20 yrs lifetime for the piping systems they design, hence cant (always) use that section of B31.3 for jusitification.

Yes and no. Could be possible that they use those paragraphs but for (a) the system must be completly identical and I see the same problems for suppliers of engineered systems which are in most cases different in piping. For (b) the comparable system needs to have any analysis.

For some equipment it is not that difficult from my point of view. Some suppliers do have the problem to consider certain values because they are afraid to make e.g. the vessel unnecessarily expensive. Some of them dont have this problem and fix just some small values to have something availabe. For vessels this should not be any problem because there are defined rules in ASME VIII or EN 13445 for considering those loads. Pumps are more critical. I Have some colleagues who are working for the pump department and they say the deformation is more critical than the stress in casing because of small gaps. In worst case the vane would hit the casing so they often say it would be better to isolate the pump from the piping with e.g. expansion joint. But they give some allowable loads. The problem is that this value are so old that nobody has any documentation for that, maybe they are just set out of experience.

I also know the problem with 0N /0Nm. I also got some colleagues who are desgining systems like that for 25 years now. What I can say is that they don't care about. Every systeme is engineered and different and it would be a high effort to check the allowable loads individually for each system. Most of them even don't have the knowledge to check this. There are some allowable loads but same for the pumps: There is no documentation and persons who fixed those values already left the company. Same for pipe stress. When you built systems for more than 25 years and there were never any problems most of the designers do not have the motivation to change anything though the spec requires it.

What I see is that it is more often required to provide any documentation for this kind of stuff to the customer. Especially when it is a big company. In most cases this appears when the systems are nearly finished in design and you have the problem to get the calculation conform specification afterwards. Sometimes you are lucky, sometimes not....