ASME / Double design temperature for a welded line

[Fran67]

Good morning everybody,

I have a question about "design temperature" notion of a piping line.

If you consider a welded line with a "T" form:
- In the main branch, the process flow is minimum -150°C
- in the perpendicular branch, there is no flow because there is a plug*.
(See the picture attached)

Because the length of perpendicular branch is 5 meter, and because there is no flow in this line, we consider that the mini temperature at the extremity of the perpendicular line will not be lower than -30°C (mini ambient temp).
The minimum permissible temperature of the welded plug is -30°C. (located at the extremity of the perpendicular branch).

Tu summerize, we have a welded line with two different minimal design temperatures

The question is: what is the design temperature of the welded line? ("design temperature" mentioned in the documentation: calculation note / test procedure...etc)
-> If -150°C, there is a contradiction with the mini permissible temperature of the plug
-> If -30°C, there is a contradiction with the process temperature (-150°C)

Do you know if a chapter of the ASME or other code deals with this case?
Or do you know simply the response?

*The plug is a special plug for this specific case of process application

Thank you for your return,
Fr67

 

[LittleInch]

No "code" I'm aware of will deal with this as it is so specific.

My concern on this is that clearly at -150 you're going to have some pretty serious insulation as indeed you need on the branch line, so I really don't think you're going to see a temp drop that fast in 5 m, but if you've done some fancy CFD models then I might believe you. Only "might".

The "safe" thing to do is rate the whole thing to -150.

Now you're trying to do something different so you need to look at the issues and document them and get them reviewed and approved because you're now doing something which isn't totally "safe".

Basically only you can determine if one end is at -150 will the temperature at the other end (-30) be capable, in any set of circumstances, from falling below that temperature. If no then just mark on the P&ID a temperature break point where everything one side is one design temp and everything the other a different design temp. To make sure you might need to add some heating elements or alarms or monitoring to kick in if the temperature of your plug gets below -30, strange that that might seem.

All the code tells you to do is do the design work and rate your components to meet the required design. It won't tell you specifics like this - this is called design.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Fran67]

Thank you for your response LittleInch,

We consider a temperature break -150°C/-30°C on the 5 meters line because there is no flow in this line, and because the process fluid is a gas.
In case of fluid flowing in the line, I agree that temperature can't decrease from -150 to -30°C with only 5 meters pipe. But for all my design cases there is no flow, never.

Maybe my example (specific) would be similar as a more general application , like a remote mounted pressure transmitter:
Sometimes pressure transmitter are mounted far from the process line to keep it away from the process conditions (when very low or very high temperature process for exemple).
In this case, there is a temperature break between the process and the PT, somewhere on the tubing line between process and PT, and this break is possible because there is normally no flow between the temperature element and the process line.
(Maybe this case is not safe or not a good practice?)

"just mark on the P&ID a temperature break point where everything one side is one design temp and everything the other a different design temp"
Do you know if it is a common practice in process industry, in case when the break is located in the middle of the line? (midlle of a line = not at a flange or fitting)?

Thank you

Fr67

 

[LittleInch]

I understand there is no flow, but there is metal in the pipe wall and metal conducts heat.

Like I said, the pipe will be heavily insulated and hence the heat coming into the pipe from outside sources through the insulation may well be enough to generate a temperature profile along the pipe. I don't know but you need to quite sure - I wouldn't be guessing or making assumptions without either some on site temperature measurements or some decent CFD modelling. You can't just say this is so because there is no flow - you need to prove it one way or another - IMHO.

If the temperature ended up being lower than your -30 and the plug failed then you would leave yourself open to criticism.

Also you may need to consider what happens if this plug leaks a little bit. Very recently I cam across the end point of a long drain line from an LPG tank which normally was at ambient temperature. Just before the valve the line sprung a small leak. Over a period of time this line then slowly froze up and ended up close to the LPG temperature and triggered a whole series of alarms and big frosting. It really doesn't have to be much of a flow to slowly freeze the line down to near main line conditions.

So long instrument lines are probably Ok because they are very small and heat transfer is low and there should not be any flow.

My recommendation would be to have some sort of temperature monitoring to confirm your analysis.

Yes you can mark changes wherever you want. In this case I would add a note and say something like - Design temp change located 0.x m away from the end flange or item.

But do the work / analysis and show to a reasonable standard that the temp won't be lower than -30C in YOUR PARTICULAR circumstance.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[danschwind]

I'm unable to wrap my mind on why a capped pipe, of only 5m length, would have such a higher temperature than the main flowing line.

I would consider it rated to -150C too so I could sleep well at night...



Daniel
Rio de Janeiro - Brazil

 

[1503-44]

That isn't even close to a viable design condition. Both the metal AND the gas will transfer heat, the metal by conduction and the gas by convection, and neither the pipe metal nor gas in that branch will remain at -30. Unless you have a heating unit at the end of the branch maintaining it at -30 temperature, all the heat contained within that branch will be quickly dissapated into the -150 branch, probably within a few minutes, if it takes that long.

“What I told you was true ... from a certain point of view.” - Obi-Wan Kenobi, "Return of the Jedi"

 

[LittleInch]

A few dimensions other than length would help.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Fran67]

Hello,

Just to clarify the application data:

When I say that the min process temperature is -150°C, it is a transient case or a failure case.
During normal operation, process temperature is higher (0°C).

Transient case or failure can not appears more than 10 minutes, because there are alarms and trip to protect it.
We consider that the -150°C of the process can not reach the plug in 10 minutes.
I think it is a safe design.

Maybe my problem is just a documentation problem, because in some documents we use (internaly or from customers), there is only one input for minimal design temperature of each welded line, and in my case there is a temperature break in the middle of the line, so there is two minimal design temperatures.
For example for a pressure test procedure in real conditions, the line must have only one design temperature: the pressure test must be at -150 or -30°C, we have to choose. (Maybe a fitting or a flange is mandatory at the break location, to split the line in two parts for two different tests, when a real conditions pressure test is mandatory).

Thank you for your opinions,

Fr67

 

[1503-44]

I don't work with temps that low, but I wonder if it is actually possible to have -150, applied in what may be a quick freeze scenario, on one side of a "plug" barrier and 0 deg on the other, without developing some thermally induced cracking. Seems like one heck of a rapidly applied thermal stress across that plug. The pipe diameter on the cold side is going to get smaller. I wouldn't want to do that thermal cycle very often. Anyway, it's your baby and you're not really telling us all that much about it. Working out the documents first sounds like that would be a good idea.

“What I told you was true ... from a certain point of view.” - Obi-Wan Kenobi, "Return of the Jedi"

 

[Ehzin]

Fran67,

Is the orientation of the deadleg vertical or horizontal? And you mention the dead leg length is 5m what's the size and length of both the run and branch line? For liquid cryogenic lines, a concept is employed on extended bonnet valves where the packing is orientated above the line and then insulation not installed past a certain point where a vapor column develops increasing the temperature at the package to help with sealing. There's a fair number of resources out there but what happens naturally is that convection is minimized if a vertical orientation is maintained. A temperature gradient is established that is continually reinforced to minimize heat transfer through the process fluid.

In your scenario, assuming the cap and dead leg is at the top, hot vapors will rise while colder vapors will orientate towards the bottom where there won't be any substantial convection or circulation in this dead leg. Temperature probes to monitor, stripping of insulation, and further analysis are recommended. Regarding the temperature probes, if this is a rare transient condition and you put probes to monitor what does that buy you when an event happens and if it fails? The temperature probes while useful will only validate this after the event happens in the future. I've seen resources that discuss this where dead legs have fairly substantial temperature gradients although I'd have to locate where I've read about that before.

I'd recommend breaking this into simpler pieces to get the individual contributions of each portion of this. To get a sense of impact of metal conduction, consider treating this dead leg such as a vessel skirt to balance ambient heat losses to find the temperature at the end of the "skirt." I believe Piping and Pipelines Assessment Guide by A. Keith Escoe has resources on determining this temperature. This wouldn't be exact by any means and there would be a lot of additional items to consider but it would at least give a best case temperature that the cap would see. If it doesn't pass in this scenario then all other considerations would only make it worse.

In any case, it's recommended to replace the cap with an appropriately rated component at the next possibility. While validating this is an option it wouldn't be recommended for a long-term strategy. Maybe knowing what this "cap" is would also help, is this simply a welded cap? I may have missed where this was mentioned. There's a worthwhile discussion to be had on this. If I find more resources I'll provide an update. Anyone, feel free to provide commentary on the above.

Thanks,
Ehzin

 

[weldstan]

You appear to be selecting the -150 design based on Joules-Thompson cooling during a very high depressurizing event of the gas at the 0 degree operating temperature. Why do you not think that the gas in the stagnant dead leg would not also depressurize as well?