Dear all,

What type of restraint is to be used on a rigid flange type connection at compressor suction and discharge of the inter stage suction and discharge connections of the compressor. If a fixed anchor point is used, what should be the stiffness value of the anchor point?
If any one knows any information on VDI stiffness values, and how they are calculated please let me know.

If the manufacturer cannot provide it, and it cannot be calculated in NozzlePro, then I would be inclined to not treat it as flexible. It would be difficult to qualify a response without knowledge of the geometry and material.

Hi Chasndrakant,

Just curious, what does VDI stands for? Try to search it in google but cannot find it.

Anyway, if it's reciprocating compressor then you might be connected to pulsation bottle which can be treated as vessel then you can use stiffness values taken from FEA. If it's centrifugal compressor then usually the vendor will give displacements at nozzle connection so you need to input that as displacement boundary condition. Usually stiffness value is not given for centrifugal type. That is based so far on my experience.

Warm Regards,

Hello Borzki,
VDI is an association of German Engineers:- http://www.vdi.eu/. Probably you can know it better if you take a look at the link I've pasted up.

For your reply please refer the following points:-
1.) The analysis I'm doing is for Inter stage piping of a 6 stage centrifugal compressor.
2.) The compressor manufacturer has not provided me with the displacements at the suction and discharge nozzles of any of the stages
3.) I've modeled each stage of piping along with the inter coolers separately and then combined them using "include piping files" & defining the global co ordinates of the Inter coolers
4.) I've modeled the suction and discharge nozzles of the compressor as individual anchor points.
5.) As per default value of CAESAR-II the anchor point has a stiffness value of 1*10^12 N/mm. This value is not realisitc
6.) If I perform a calculation using this value the magnitude of imposed loads on the nozzle of the compressor is very high and it will fail in API 617 and NEMA SM-23 analysis.
7.) For this reason i need the realistic value of the stiffness of the suction and discharge nozzles of the compressor of each stages
8.) The VDI has done some work in calculation of the stiffness values for various flange sizes. i wanted to know them. But apparently no data is available on internet about it.
9.) Could you please provide me your suggestion weather my modelling strategy is correct or not. If required I can send you the CAESAR II file that I've modeled.

Do you have any sort of detail that might indicate the wall thicknesses on the casings? I would envision it to be pretty thick, and I suspect that were you to run a NozzlePro type analysis, you would get stiffnesses in the range of 10^6 lb/in to 10^8 lb/in for 1" walls.

I would test at unrealistically flexible connections. If you can't get it to pass at something that you know to be unrealistically generous, then you know that you need to fix the piping configuration.

For all other configurations, your only alternative is to model it in such a way that guarantees the nozzle loads to pass muster, regardless of how flexible they connections might be and how inflexible you assume them to be.

You don't know what you don't know, and if the manufacturer refuses to assist, then your only way out is to cover yourself through robust design and with listed assumptions.

Hi Chasndrakant,

Thanks for the VDI info. Michael is correct, it's hard to assume a stiffness value in the compressor nozzle/casing interface as the geometry of compressor casing is not straightforward unlike tanks, vessels or exchangers where there are readily available template based FEA software packages to extract stiffness values. It will be a tedious task to extract this stiffness values for compressor. I haven't tried this approach also in my previous experience.

What I can suggest are the following (depending on the phase of the project such as Tender, FEED or EPC stage):

1. Normally in our company we set standard nozzle loads in collaboration with the Mechanical group such as the following:

2 x API 610 for API pumps (centrifugal type).
3 x NEMA SM23 (for centrifugal compressor)., etc.

The Mechanical group will use this info to be given to Vendor as their basis. In this way, in the upfront stage of the project the Vendor can already design their equipment based on this nozzle loads.

2. Also we request the displacements at nozzle connection for different operating cases of compressor as part of Vendor's deliverables.

This is applicable if the stage of the project is during Tender or FEED stage.

But if the stage of the project is at detailed engineering EPC stage, usually the compressor vendor will ask for additional cost if you request higher allowable nozzle loads than the standard. In this case, you can talk to your Piping Lead Engineer, Mechanical Engineer and Project Engineer to choose which is the most economical solution.

Usually for centrifugal compressor, a strut is used depending on the piping configuration and displacement at nozzle connection in order to nullify if not minimize the nozzle loads (thermal) to meet the criteria.

Please share to us what solution you have come up with. Maybe that VDI approach (I haven't look through this document yet) can be a more viable solution.

Any other opinion is greatly appreciated.

Cheers!!