For an earlier executed project for pipe stress analysis of 5 stage centrifugal compressor,
We have some queries raised by our end client. We would like to get them sorted out and
also increase our understanding about correct modelling techniques in CAESAR.


1.)We have modeled the inter stage coolers (Heat Exchangers) as a pipe element filled with water. Is this procedure correct. If yes / no can you please explain.

Also in standard EPC practice heat exchangers are modeled as weightless rigid elements. Is there a specific
reason for the same?

2.)Our client has commented us to model the Heat Exchanger as a rigid element with weights assigned to those rigid elements. Is this procedure correct. Please provide explanation.

3.)The Heat Exchangers are supported using spring hanger support instead of using saddles. Also we do not have any input data for spring hangers. For such situations how can we select
a correct spring and provide inputs to get correct spring designed for the system.

Ultimately, you are responsible for modeling the appropriate movements and reactions of nozzles and keeping them within allowable loads of the nozzles (which may be validating stresses of equipment), and within allowable stresses of the pipe.

There is no one technique that covers all kinds of heat exchangers, so you'll need to be more specific with your request.

Hello Michael,

Request you to please reply serially to my queries.

As far as keeping the stresses within allowable limit are concerned, please note that this is a 5 stage centrifugal compressor. The team at Hexagon / Intergraph are yet to come up with a solution to an approach where in an inter stage piping and nozzle load analysis of a multi-stage centrifugal compressor is required. The application engineering expert them selves have said to me that they have yet to come up with a module in CAESAR-II where in such an analysis could be performed.

Also based on our firm's study we go along with an approach where we model the piping as per geometry of the inter stage piping. Then the flanged ends of the pipe which are coupled to various nozzles of the inter stages are anchored. These anchor points are assigned with finite stiffness values as per VDI stiffness. Then the results are obtained for such model.

We then compare the global loads obtained for the anchor points with the equation of clause 8.4.6.1 of NEMA SM-23 or Appendix-2E of the API 617 to conclude whether the imposed loads on the nozzles are within their allowable limits.

The heat exchanger used in our particular case is a Shell and tube type heat exchanger

Generally, this is acceptable for typical pressure vessels, such as shell and tube heat exchangers. However, this is only a partial picture.



Generally, filling a vessel with water is an acceptable practice.

However, your vessel is on springs. If you specify a vessel weight, spring rate, cold load contrary to reality, then the accuracy of your model will suffer.

Your task is to model actual nozzle displacements.


When you specify rigid in CAESAR, you're only saying that it's a piping element that's 10x thicker than your inputs, and increasing any insulation weights assigned to that item. Generally, you can get away with assigning rigid status to any equipment so long as the equipment is in reality much stiffer than the piping you connect to it. Regardless how you model the body of the vessel, I recommend modeling the nozzle as pipe.



See earlier explanation. To add to that, you generally should model to the vessel centerline. Offset nozzles need a pathway modeled to that centerline, and you generally use rigids to get to the centerline.



Offer to size new spring hangers.

Alternatively, offer to size how the hangers "should" have been originally sized, and use those values, instead.

Either way, you'll need end-user approval on how to proceed.