How model a line for stress analysis 3

[flash104]

Hello all,
I've just started to work on stress analysis.
I have to add 2 valves 12"-150# to an existing line (12" fluid:hydrocarbons)and calculate with B31.3 lines.
Now, I have 2 problems:

1) This line is very long and has a lot of ramifications and the anchors are very far from interest zone.
Have I to model ALL the line or i can stop in some specific point?

2) In which way can i model a three way valve?

I'll use Algor Pipeplus.

Sorry for my criminal english.....thanks

PierAndrea

 

[pipesRus]

To my knowledge there is no effective way of de-coupling the majority of your model without using a full anchor. There are some techniques but it depends on geometry and the loading conditions, the most effective way is the anchor.

Not sure about Pipeplus, I'm a ADLPIPE user. In ADLPIPE a three way valve can be created with three separate valve instruction having a common node where the three legs meet.

Phil

 

[flash104]

Thank you pipeRus.
I also think that, but the client doesn't want to "buy" a 200 ft stress analysis for adding two valves.

I might elect some points like "end point calculation " and input their displacements.

PierAndrea

 

[NozzleTwister]

Pier,

Are you added a branch to the existing line with the 3-way valve or are you just dropping in a couple of valves to the existing line?

If you are just adding valves and no additional piping you may not need to do stress analysis if your valves will be supported adequately from the line and there are no spring supports affected.

If you're adding a branch that's a different story. What's the temperature? Is a flexibilty study required?

If you do need to do a stress analysis, the boundry conditions of your problem is very important to achieve accurate results. If your pipe has piping has some changes in direction and is guided you may not need to go back to the anchor. If the piping is in a rack, the extra coding should just take a few minutes. If you select your boundry by inputing displacements, be sure your number is reasonably accurate and not very close to you branch. A guide is a good place to do this. You just want to make sure the point you choose does not affect the flexibilty of the system at your area of concern. In addition to normal operating temps, also consider steam-out and upsets as applicable.

Good luck,

NozzleTwister
Houston, Texas

 

[StressGuy]

I'm reading the description to suggest that you are adding a set of valves in-line in an existing pipe run. If so, you should be able to determine by observation if a formal stress run is going to be required by seeing how the valves sit in relation to the rest of the routing.

If you've got a long rack run with typical 20ft spacing between supports, I'd doubt you'd need to make a model, just make sure the valves are relatively close to the support crossing.

If they're in the middle of an expansion loop, you would need to look at this very carefully. The rigid bodies reduce the effective flexibility of the loop. Plus, since loops are, by design, higher stress areas, you would have a real concern that bending at the flanges (assuming typical flanged valves in your application) could create leakage.

As for a three way valve, typically you would model it with three rigid elements with a common node at the point where the three connected pipes intersect.

Edward L. Klein
Pipe Stress Engineer
Houston, Texas

"All the world is a Spring"

All opinions expressed here are my own and not my company's.

 

[flash104]

Nozzle, Stressguy,
this is the existing situation: the line runs on a rack (elevation ~12 ft), then comes down to a control valve group, then rides up to the rack again.
I have to add a valved bypass for the control valve and a block valve.
There’s no guides, only supports.
I think it’s only a weight problem, maybe analysis is not techincally required, but the company has to certificate every changes…

Thanks a lot for your hints, you’re teaching me a lot.

 

[PAN]

flash104,

Computer analysis may be not necessary.

Is this line in the high temperature service condition? You may require only checking piping flexibility by visual inspection.

 

[flash104]

Pan, the design temperature is 210°C (410°F), the operating temperature is 177°C (351°F).
Maybe it will be necessary to arrange a support under new valves.

 

[StressGuy]

Let me see if I have the picture correctly - I'm taking your control station to be a typical setup where the line comes down, you go through an isolation block valve, go horizontal through the control valve, and then back up vertical through a second block valve. I've also seen it with all three valves in the horizontal.

As I read it, you are adding a connector with block valve between the two verticals above the block valves so that you can have flow through the bypass while the main block valves are closed off (to service the control valve, for example).

If this is the case, you do have a concern, particularly at this temperature. Putting a bypass straight between the two vertical will stiffen the piping system quite a bit. Most of the control sets I've seen are setup with this bypass arrangment initially and if there's a stress problem, is it almost always at the branch connection of the bypass. A good stress engineer won't use a control station as an expansion loop, but if your anchor points are as far away as you describe, that may be exactly what you have.

Now, the other way that I could read is that you literally have one block valve and one control valve in series and you want to put two bypasses, one around each valve. In that case, I would only see it as a support issue.

Edward L. Klein
Pipe Stress Engineer
Houston, Texas

"All the world is a Spring"

All opinions expressed here are my own and not my company's.

 

[recoveringEngineer]

StressGuy
Good post. I found this entire thread interesting.
I pictured the arrangement as your first sugestion, also. I am not posing as an expert in this forem but do offer this for discussion purposes:

Certainly, closing up the loop with a straight through section would severely stiffen the run. If that was the required flex for expansion it would be unacceptable. However, the bypass and block, also need to be accessible which may mean, they can be at grade in the horizontal plane and not "severely" stiffen the loop.

But the flexibility issue needs to be addressed.



Donald Blachly, PE

 

[flash104]

Stressguy,
Your first suggestion has correctly described the situation and your post is really interesting.
I think flexibility is not place in risk, in fact the line has a lot of change of direction( crosses the rack many times, change elevation....).
However i agree with you, hazardous situation is the branch connection of the by-pass.

Yesterday i inspected the plant, it’s a Topping, and i saw a lot of control station "already by-passed" in rack running pipe.

I wasn’t sure to work like stress engineer but now I am becoming fond of these subjects, thanks guys!

 

[DSB123]

flash104,
I disagree with you and beleive flexibility could be an issue. You have to consider the thermal case of the two block valve either side of the control valve being closed with the flow through the bypass. In this situation the bypass is hot whilst the control valve leg is cold. If you have insufficient flexibility in the bypass leg or the leg is situated close to the main run then high stresses can develop at teh branch connections. Suggest you review the "local" flexibility around the control station.

 

[flash104]

DSB123, you are actually right.
I was thinking to connect the two vertical lines with a "C modeled" leg, (likely an expansion loop) with vertical lines in the C tip and the valve in the middle.

Now my biggest issue is to determine boundry conditions.