AutoPIPE Wiki 12. How can I insert the shell connection on the head of a heat exchanger in AutoPIPE Vessel?

12. How can I insert the shell connection on the head of a heat exchanger in AutoPIPE Vessel?

Applies To
Product(s):	AutoPIPE Vessel
Version(s):	V8i
Environment:	N/A
Area:	Modeling
Original Author:	Bentley Technical Support Group
Date Logged & Current Version	August 2016 40.02.00.08

Problem:

How can I insert the shell connection on the head of a heat exchanger in AutoPIPE Vessel?

Example:

How can I insert the shell pos. 11 in the head.

Solution:

This is a special configuration of heat exchanger for which there is no standard template available in AutoPIPE Vessel. You may model the exchanger in 2 parts:

Part #1

Model as a BKU exchanger, this will calculate the shell, channel tubeside, flange, tubesheet etc.

Part #2:

Create a pressure vessel model with same shell geometry, shell side design conditions of heat exchanger and insert an offset nozzle on the ellipsoidal head which is assumed as a manhole with a bolted flange cover. This will code check the head (item 02) + nozzle (item 11) configuration. I modeled the nozzle as self reinforced to simulate the geometry and specified the reinforcement height, outward and inward projections etc.

Another approach, previous mentioned using a manhole nozzle, suggest that instead of a manhole nozzle consider modeling using a regular process nozzle. A manhole nozzle does not accept nozzle local load. However, one can apply the nozzle local load which is equal to the vertical loads and moments exerted from the tubebundle (if it is NOT supported by baffles or maybe some partial load is exerted due to the weight of the bundle at the nozzle flange) to do a more conservative analysis.

The idea is if the nozzle is seeing any moment or shear load then for a realistic model should consider that load being applied at the nozzle to shell junction and perform a local load analysis based on any of the methods available, which is your choice (eg. WRC, PD5500, EN etc.).

Consider also performing an FEA analysis if you see the dimensions exceed the applicability of the local load methods.

Note,

1. The approach of FEA and WRC are quite different. Nozzle local load analysis is a complex engineering problem and there is no exact solution. WRC is a closed end approximate solution with limitations of application and FEA is a numerical method and likewise not the exact solution either and the results of WRC may vary from FEA, it is normal. Therefore, the pass/fail depends on what criteria is considered for acceptance.

2. The above suggested method is just for reference, final design modeling approach is a decision of the designer and must involve sound engineering judgement before making any modeling assumptions.