Refinery Fluid Thermal Expansion Due to Steam Tracing

[KCAB14]

Hello, i am currently a process engineer at a refinery and im looking into the effects of steam tracing on lines and the resulting thermal expansion if the line is blocked in. The lines are full of hydrocarbons and I was wondering if line length matters as we typically only see leaks on our longer lines. Me question is; if a one foot line is blocked in and steam traced at 350 oF and a 100 ft line is traced at the same temperature with the same fluid inside and fill volume percentage, will the resulting pressure on the pipe and flange be the same, or does total heat into the system matter at all? Hopefully this makes sense, im essentially trying to figure out why we are having more leaks on our longer lines than our shorter ones.

 

[EmmanuelTop]

This subject (or related to it) has been discussed recently at

http://www.eng-tips.com/viewthread.cfm?qid=389919

In theory, there is no difference between long and short lines if the same heat flux were applied. In practice, as you have observed yourself, the differences do appear - mostly due to localized overheating, condensate pocketing, etc.

For longer lines, installation of TSV is recommended. Various companies have different guidelines with regards to the liquid volume threshold (volume of trapped liquid above which thermal expansion relief needs to be provided), and this is in line with field observations that shorter lines usually do not experience overheating. You can read more on thermal expansion in API 520 and

http://www.chemwork.org/PDF/papers/what you should know about liquid thermal_expansion.pdf

Dejan IVANOVIC
Process Engineer, MSChE

 

[bimr]

If the fluid in non-compressible, it should not make much distance as to the length of pipe. Trapped liquids can generate large pressure increases with only modest temperature changes. The phenomenon is understood by the process safety community. However, it is also a phenomenon that is easily overlooked or dismissed as being of little to no consequence.

http://abiquim.org.br/congresso/cong_cd/fullpapers/p171757.pdf

 

[BigInch]

Longer lines will have the same percentage increase of fluid volume (as the fluid makes its way out of the pipe) but of course the net volume increase will be greater. With a greater volume potential to escape, more paths out of the pipe will usually be found, and with the longer pipe, there are typically more paths for the fluid to take. There are more leaks on a longer line, simply because there are typically more possible paths to open up (ie faulty girth welds, flanges, tees, valves, corrosion pin holes, etc.).

 

[KCAB14]

Thanks for all your input I appreciate it. Yes i know that in theory with the same applied heat flux the pressure should be the same but its just weird how it always seems to be on out long lines, maybe it is something to do with more fluid paths. I know that with our naphtha its mostly gas at steam trace temperatures when we have had a leak. But we have also had a leak on our gas oil which doesnt even start boiling till 600oF so its just liquid expansion. Im assuming this would still be enough to cause a leak but why then never on our small lines... its a mystery to me. I know there may be a theory as well but id like to be able to prove it mathematically but i dont think this will be the case.

 

[SNORGY]

I think the other thing that comes into play is that valves typically leak across their seats, irrespective of whether or not API allows you to take credit for leakage towards the determination of the requirement for a TSV. So, systems that theoretically should leak, don't. Well, they do, but they stay inside the pipe.

 

[georgeverghese]

Thermal expansion relief varies directly as the heat input (from heat tracing) per unit mass of blocked in volume, asssuming specific heat and cubical expansion coeff is about the same and there are no leak paths.

The TSV leak rate can be calculated if you know how the liquid density changes with temperature rise; usually mass rates are low and may be accomodated with a 1D2 orifice or even smaller, but if you face the risk of a engineering safety audit, better to have these rates calculated.