How to Manage Pipe Support Lift-Off in Stress Analysis

[nikolastrojman]

Hello everyone,

I’m seeking your opinion on managing support lift-off scenarios, particularly in systems where there’s a significant difference between "normal" operating temperature (e.g. 60°C) and design temperature (250°C).

Here’s the situation which I encounter many times when doing stress analysis:
- At ambient (installation) temperature, the rigid support carries the pipe's weight as expected
- At normal service temperature, the stress analysis software shows vertical displacements in the range of 0.5 mm to 2 mm
- for design temperature, the lift-off becomes more pronounced

Should we consider only the design temperature scenario and use variable pipe supports/hangers, which are very expensive and are treated as specialty support items to be used only when required -> looking from practical/economical point of view.

Let's assume hot-sustained has already been checked and stress are in compliance with the code (the system is also ok without the supports which are lifting).

We are talking here only about whether to leave the rigid support in place, remove it (which increases line sagging then) or switch to variable supports.

Which are the strategies you use to avoid overengineering the system from support concept point of view.

 

[XL83NL]

The Code isnt a cookbook. Review some more details on things like thermal shakedown in e.g. Peng or Maan & Jawad (both in ASME Press series). It'll help understanding these concepts better.
The Code recognizes this also. Local yielding, that may occur in situations where small amounts of liftoff are present (which will always be there as no new system is perfect prior to startup), will diminish over time provided the liftoff is within an allowable range.

 

[RVAmeche]

David,

So in Caesar II that's handled via their "alternate sustained" cases - these are basically sustained stress checks based on your operating case so if you have support lift off scenario, you'll see the increase in sustained stresses in those cases.

The previous general discussion of engineering judgement of when to use a spring vs rigid support is still valid - if you go "strictly by the book" any hot system could result in a ton of spring hanger supports because everything's moving. This isn't economical nor a good design. Naturally when reviewing what supports to eliminate or change, you should always be checking sustained loads and deflections to make sure it's acceptable.

 

[David2007]

David,

So in Caesar II that's handled via their "alternate sustained" cases - these are basically sustained stress checks based on your operating case so if you have support lift off scenario, you'll see the increase in sustained stresses in those cases.

The previous general discussion of engineering judgement of when to use a spring vs rigid support is still valid - if you go "strictly by the book" any hot system could result in a ton of spring hanger supports because everything's moving. This isn't economical nor a good design. Naturally when reviewing what supports to eliminate or change, you should always be checking sustained loads and deflections to make sure it's acceptable.

Thanks for your feedback, I agree with you that it’s not economic to use spring support for all lift-off cases, however my point was to not to ignore any lift off support without engineering justification, and the formula you have provided to calculate the distance between the first spring hanger and the rigid support can optimize using of spring support.

In the end, can you help me to understand your statement ‘if you have support lift off scenario, you'll see the increase in sustained stresses in those cases.’ , since I think the lift off takes place in operating/Thermal load cases.

 

[RVAmeche]

Previously you may have had something like:
Case 1 = Operating Case (support lift offs but no allowable stresses)
Case 2 = Sustained Case (as-installed, no lift offs)
Case 3 = Expansion Case (total stress range, just used for allowable stresses, no lift offs)
Without careful review of Case 1 you may miss the support lift offs and excessive sustained stresses in that condition.

The Caesar II "alternate sustained" cases are tied to a specific operating case.
Case 1 = Operating Case (support lift offs but no allowable stresses)
Case 2 = Alternate Sustained for Case 1 (determines sustained stresses based on support configuration in Case 1)
Case 3 = Sustained Case (as-installed, no lift offs)
Case 4 = Expansion Case (total stress range, just used for allowable stresses, no lift offs)

Here Case 2 = "alternate sustained" associated with Case 1, so it's looking at sustained stresses based off the active supports in that operating scenario. If you had 5 supports lift off in a row, you'd see that alternate sustained case fail code stress. This is satisfying the "all support conditions" sustained stress statement of B31.3 302.3.5 in later editions of the code.

 

[David2007]

Previously you may have had something like:
Case 1 = Operating Case (support lift offs but no allowable stresses)
Case 2 = Sustained Case (as-installed, no lift offs)
Case 3 = Expansion Case (total stress range, just used for allowable stresses, no lift offs)
Without careful review of Case 1 you may miss the support lift offs and excessive sustained stresses in that condition.

The Caesar II "alternate sustained" cases are tied to a specific operating case.
Case 1 = Operating Case (support lift offs but no allowable stresses)
Case 2 = Alternate Sustained for Case 1 (determines sustained stresses based on support configuration in Case 1)
Case 3 = Sustained Case (as-installed, no lift offs)
Case 4 = Expansion Case (total stress range, just used for allowable stresses, no lift offs)

Here Case 2 = "alternate sustained" associated with Case 1, so it's looking at sustained stresses based off the active supports in that operating scenario. If you had 5 supports lift off in a row, you'd see that alternate sustained case fail code stress. This is satisfying the "all support conditions" sustained stress statement of B31.3 302.3.5 in later editions of the code.

Thanks for the clarification.