Cold temperatures upstream of valve 2

[will48]

Does anyone have a feel for (or ideally some reference to) how far UPSTREAM cooling will take place in the pipe wall at a valve in gas service with a high dP.

In other words, a hydrocarbon gas at ~140 bar is depressured to flare (atmospheric + nominal back-pressure) via a globe valve. Valve and downstream pipework fabricated from Low Temp St.St. Upstream pipework is carbon steel whose minimum design temperature is greater than the expected gas temperature downstream. Should we specify a length of LTSS upstream of the valve also to account for conductive heat loss?

Thanks
Will

 

[Montemayor]

Will:

Get serious. No one can tell or predict "how far UPSTREAM cooling will take place" in YOUR pipe wall until you tell us the heat transfer taking place and at what rate. We can't tell; only you know the local conditions and data. The pipe could be fully insulated and the flow nearly adiabatic. We can't tell.

If you are truly realizing downstream cooling due to free expansion of a hydrocarbon gas, then you have reason to be concerned about the metallurgy of the downstream piping and equipment. But is this presently occurring? Or is this operation in the design stage? Be specific. If it is operating, then the answer exists in the field. If it is in the design stage, give us all the design basic data. Fair is fair. We're here to ty to help; but you've got to help us help you.

Additionally, the expansion cooling will take place DOWNSTREAM of the globe valve, not UPSTREAM. The globe valve itself should be within the pipe specifications for the lower temperature operating condition. Under flow conditions there will be little or no conductive heat transfer upstream of the glove valve. If you want to be supersafe, allow for 3-6 ft of low temp pipe upstream of the valve. That's the best I can recommend with the basic data supplied.

 

[will48]

Montemayor

I will forgive the abrupt opening to your response on account of your long history of politeness and patience. I was merely asking for an order-of-magnitude feel, not a transcript of the calculation, hence the lack of data supplied. Two of your final three sentences were sufficient, and much appreciated.

Regards
Will

 

[sshep]

Will48,

I have felt a lot of piping to compare upstream to downstream temperature for flashing liquids as well as sensible vapors, and can tell you that the cold temperature does not extend upstream very far- essentually not at all, but at most a few diameters for uninsulated pipe.

Yesterday I felt for some flashing pentane in a 2" line and while the downstream and control valve was cool, there was no discernable effect upstream of the valve flange. Between the insulation effect of the gasket, heat transfer to warm inlet, and losses to ambient, there was no hint of what was happening downstream.

Even in cases of flashing liquid where ice formation tells the story visually, the upstream icing is typically inches. This is because the cross-sectional area of metal in a pipe is small relative to the surface area- i.e a 4"pipe is 0.002sqm cross-section of metal. If you do some simple and quick (Q=k*A*dT/t) longitudinal heat transfer calcs vs circumferential heat transfer to process and ambient you can quantitatively convince yourself that very little upstream distance is needed.

best wishes, sshep

 

[hacksaw]

wil 48,

chances are that you have a spec break at the valve. if you have a block and bypass arrangement, most of the cooling would be down stream of the hp block.

to be safe take your worst case cooling condition, and allow for 2-5 d of upstream, good luck

 

[25362]

If the depressurizing is done from a closed vessel, the pressure reduction in the vessel would most probably cool the gas upstream the valve.

 

[sailoday28]

I am not a chem engineer, an unfamilar with the use of "flares" However, looking at the original question----
In other words, a hydrocarbon gas at ~140 bar is depressured to flare (atmospheric + nominal back-pressure) via a globe valve.-----and the response of
25362 (Chemical), the original question makes sense to me and can be analyzed.
A worst case scenario would be neglecting heat capacity of the upstream piping and assuming no heat transfer.
Further, the upstream piping does not have to be a closed vessel, just that the valve in question is initially blocked. When the blocked valve is opened depressurization occurs and eventually a steady state is reached.
Am I missing something with regard to a flare?
Regards

 

[MortenA]

I will suspect that the reason why the "cold" does not "creep" backwards is that you have a continious stream of warm fluid flowing towards the valve. This will heat the metal and the "cold" can thus not "creep" backwards.

Best regards

Morten