B31.4 Expansion Stress understanding

[Fidens]

Hi all,

I'm working on the stress analysis of a pipe rack, per B31.4. My question is based around the code requirements for expansion stress and correctly reflecting that requirement in the stress analysis load cases. From reading B31.4, section 402.5.2 (unrestrained pipe) it shows the equation for calculating the stress range resulting from thermal expansion and a note: "thermal stress should be calculated for the range of minimum and maximum operation temperatures"
If my T1 is my max and T2 is my min then:

1. Is it reasonable to assume that ambient temperature is my pipe metal temperature? Meaning, in the winter the average low is 10F and summer average high is 90F. As a result, my worst case scenarios are these temperatures(T1, T2) for my stress model. I'm aware that this approach neglects the fluid's heat transfer to the pipe during winter and vice-versa during summer (gasoline, expected to be around 68F) hence why I'm asking/confirming. Is that correct? If yes, then...

2. After creating my OPE cases (L1: W+T1+P1 & L2: W+T2+P1) I can segregate my full expansion stress range by doing L1-L2. This is saying that I'm taking the stresses calculated at min temp and adding them to the stresses calculated at max temp and getting this total "stress range". Is this understanding correct? If yes, then it complies with code requirements and the load case is correct. Yes? No?

Thank you in advance.

 

[BigInch]

Use the installation temperature as the initial temperature T0 then analyze two separate temperature excursion cases from that.

1) Minimum Temperature - Installation temperature, (T2-T0), analyze for maximum tensile thermal stresses.
2) Maximum Temperature - Installation temperature, (T1-T0), analyze for maximum compressive thermal stresses.

 

[Fidens]

Thanks BigInch ! that makes a lot more sense to me. A stress check swing from Min to Max seemed a bit too extreme too me and unrealistic.
I see that by your suggestion the "range" is more confined to operating within a "season" as opposed to the whole year which is far more realistic. Not sure if my little summary of understanding is 100% correct (I hope I don't confuse others) but at the very least your suggestion is clear. Thank you again.

 

[BigInch]

There is no induced thermal stress when the system is at the temperature of final weld in. That's your zero point from which you calculate thermal elongation/contraction that generates thermal stress, if restrained.

If not restrained, the pipe will simply expand without generating stress. If the pipe is 75% axially restrained, it will generate 75% of the full axial thermal expansion stress, but will expand, or contract, only by 25% of the calculated full axial deflection.

 

[XL83NL]

Doesn't B31.3 (I know, different code) requires the designer to evaluate thermal displacement stresses over the complete range, i.e. from T1 to T2? If so, why does B31.4 deviate from this requirement?

 

[TGS4]

One of the reasons to calculate the entire thermal displacement stress range is to compare to the ratcheting limits. Piping and pipelines are the mechanical equipment where thermal ratcheting has most frequently been observed. Its more than just a tensile and compressive stress limit.

 

[BigInch]

B31.4 and .8 are different. Ratcheting in pipelines is not common, ambient temperatures are usually not extreme, pipelines go years at a time before being shut down, and, once the pipe has moved, if it does at all, soil typially and quickly locks it in place.

 

[DSB123]

XL83NL,
You are correct that is the correct thermal stress range which needs to be calculated.

 

[saplanti]

Please do not forget that the stress ranges in B31.4 and B31.3 are different. BigInch is correct for the nominated code.