modeling compressor nozzles - static analysis according with Nema SM23 without taking into consideration gravity (only pressure, temperature loads, and gas gravity) ?

Hello Bob, 

Suggest that you model your piping up to the compressor location but do not model an Anchor or nozzle element for the compressor it self at this point in time. Analyze the model, using the results grids copy the exact displacement and Rotation of the node point location representing the compressor.

Note: By clicking in the grid fields see the exact value used by AutoPIPE in the calculations instead of the rounded off value in the output report..

Insert an anchor to represent the compressor connection, then insert Import Support displacement on the anchor, using the values recorded above for Translation X,Y,Z and Rotation X,Y,Z, run your Analysis and check the Equipment sub-report. Loads due to gravity should be fairly low, if not Zero.

See example attached model to prove concept. If remove anchors A02 and B02, The gravity displacement = the impose displacements. 

Rotating equip-pump.zip

Note: attached model created by AutoPIPE 12.02.00.14

Regards,

Mike Dattilio
Bentley Systems Design Analyst
Design Engineering Analysis group
===================================================

Hello Mike,

Thank you for explanation.  After inserting an anchor to represent the compressor connection, and importing Support displacement due to gravity load on the anchor, I can run run static analysis using gravity, thermal, and pressure loads (all loads acting on the piping). As a result  the  Loads due to gravity should be fairly low, if not Zero. I understand that; however in reality the compressor anchor is modeled as a rigid support that is not  imposed displacement due to gravity. Neverthless, by using this method we make our results acting on compressor nozzle more realistic. This is because of two opposite reasons:

1. Modeling the compressor nozzle as rigid in reality is a very conservative approach, particularly for comparable small size compressors.

2. By imposing displacements due to gravity load on anchor  in inreality is some kind of liberal approach, particularly for not well supported piping system. It is because the compressor nozzle will not move exactly as the imposed displacement of piping end (point of connection with anchor).

Considering two above reasons, and using the methodology of compressor anchor modeling you pointed out, I think we compromise (optimize) the load results acting on the compressor nozzle, in order to make them more realistic. Please correct me if I'm wrong.

Regards

Bob

Bob,

Assuming you don't want to take into account GR, I believe you may define the appropriate Analysis Set, with Gravity abandoned. Then you may report specific loading acting on nozzles.

However, discussing nozzle loading I think it would be much safer not to abandon them at all, or I do not understand clearly the system configuration. Yes, it may lead to more conservative results, however, I believe that thermal expansion generates much higher loading and this is a real challenge in this kind of gas compressor stations.

I was asked by Bob to comment more so here I am with my opinion. I think that the solution proposed by Mike may be difficult to apply when it comes to inserting such small values of a free pipeline end displacement. Values of this order can also be "method errors" especially when it comes to non-linear analysis. I am opting for the variant I propose. Please, take into account that in case of non-linear analysis, the initial conditions for thermal analysis are the results of gravity analysis (in the assumed sequence GR > T1 > P1; the order of P and T can be changed). We have a discussion regarding 'model splitting' somewhere else in the community and that subject seems to be not quite clear and solved as well. Topic open...