Stress Analysis - Lateral Bowing in Straight Pipe 1

[simon17]

Hello all,

First time poster with a stress analysis inquiry.

My company is working on analyzing a 4" line @ 150° in a tank farm. The line runs through berms and spans approx. 200' in between each set. The client does not want any loops in the pipe between the berms and would like to know the lateral deflection due to bowing. Using arc and triangle principles, it is possible to calculate the theoretical lateral deflection due to a single bow (35" for A106B pipe @ 150° spanning ~200ft).

After extensive searching online, I haven't been able to find any literature or equations to calculate multiple bowing (snaking) in a Carbon Steel pipe. Any help with this would be greatly appreciated.

Thanks,
Dmitriy

 

[pennpiper]

dso88,
This statement suggests a number of things to me.
"The client does not want any loops in the pipe between the berms"


1). The Client is correct. He does not want any loops in the pipes between the dikes because there should not be any straight runs dike to dike inside a dike area.

2). The pipe routing is incorrect. All pipe runs (to and from tanks) should connect to the tank on one end and exit the dike as soon as is practical. The Dike penetration should be via a Sleeve that is open (to drain) on the tank end and flanged off (sealed off) on the out side of the dike. The pipe run from the tank to the dike should have an "L" or "Z" configuration that allows for thermal expansion/contraction caused by solar or process heat.

3). You should never have pipes from one tank area run through the dike area of another tank. The use of Dikes is normally related to the storage of flammable hydrogen fluids. If you have a tank leak and a fire inside a dike area and there are lines to or from other tanks running through the fire you run the risk of spreading the fire to other tank areas.

Why are pipes running through dike areas?


prognosis: Lead or Lag

 

[MJCronin]

Check out this image of a Tank Farm:

http://www.airphotona.com/image.asp?imageid=7287

 

[pennpiper]

MJCronin,
The posting of "A Tank Farm" picture is not the same as posting correct Tank Farm standards.
There is so much wrong with what appears in that picture that I would never have posted it.
- Tanks too close together
- Tanks too close to roads
- No Dikes between very large single tanks
- No Dikes between groupings of four or less small tanks
- Main runs of piping too close to Tanks

However in the defense of what is shown in the picture I would offer the following:
- We have no information as to what the commodity is in these tanks
- We have no information as to when this tank farm was built

I would never try to use this as an example of the correct manner to layout tanks and sleeper pipe racks in a Tank Farm.



prognosis: Lead or Lag

 

[pennpiper]

A Tank Farm article worth reading.

http://www.arabianoilandgas.com/article-5870-petrochemicals_focus_storage_tank_farms/#.UR1aX2ecXlk

prognosis: Lead or Lag

 

[MJCronin]

Penn,

With all due respect, I did not mean to suggest that the configuration in the tank farm was correct or even optimum.

I simply thought that it would move the conversation along.

Your tank farm article is informative, but, in some ways, a bit trite:

(e.g. “All the tanks are surrounded by a dyke wall. In case of a major disaster, the chemicals can be contained within the dyke wall and there is no free run,” he adds")

I would be interested in a tank farm picture that you believe is better practice.

I have purchased your informative and well-written book and have suggested it to others.

Best Wishes,

-MJC

 

[pennpiper]

MJC,
Your comment:
"I would be interested in a tank farm picture that you believe is better practice."
brings up a disappointment I have had for years with free and Fee pictures on the web. There are lots of subjects and lots of pictures in each category but most of the pictures do not show really good engineering or design. It is easier to find bad examples of something or purely bland examples that don't show anything significant.

Being retired and out of the loop as they say, I don't have access to the resources I once had.





prognosis: Lead or Lag

 

[BigInch]

Perhaps the attachment is helpful.

For pipe with pinned ends, assume a sine wave.
You could plot points for a unit sine wave then calculate and sum the distances between each point. The difference between the arc length and the length of chord could be rationed and scaled to your particular temperature growth.

Independent events are seldomly independent.

 

[simon17]

Thanks for all of the suggestions. I thought I had it set that any comments would be emailed to me which is why I haven't responded until now.

Pennpiper, you are correct in that there should be bends to allow for thermal expansion. The client decided they'd like to have straight runs of pipe and simply let it deflect and buckle and wanted to know how large the deflections would be. We convinced them that the axial loads at the burms would be very high and the loops need to stay in. I also don't know why there are pipes running through dike areas. It's all existing piping that they are configuring.

Thank you for the photo, but it's not what I was looking for. I don't see at all how it helps with my question.

BigInch, thanks for the elastic buckling pdf. I read through it, and I also read your comment. In my calculations, I calculated the growth, added that to the total length of my run, and then used those two numbers as my chord length and arc length to find the arc height. But this is still not a real-life application, as the pipe will not simply deflect in the middle but will have multiple buckles. From all the searching I've done, there is nothing I've found that will help me approximate that.
I'm not sure what you mean by "the arc length and the length of chord could be rationed and scaled to your particular temperature growth."

Thanks,
D

 

[chicopee]

Get a book on strength of material and find continous beam equations with intermittent supports; then select a segment of piping between supports showing the rotation of the ends, end supports loads, end moments,and the longitudinal load from thermal expansions; from all of this you can determined the deflection and don't forget to apply the liquid fill pipe weigth.

 

[BigInch]

With multiple deflections, the line buckles in multiple sine waves, n* [π]/2, depending on the buckling load to strain (expansion/unit length ratio) and lateral restraint. If it buckles once and growth continues while additional lateral deflection is restrained, another buckle can form, increasing n.

For a given straight length, growth strain and change in straight length, there is only one sine wave that can be constructed, which will have one particular ratio of lateral deflection at the mid point to the thermal strain, the change in length/original length. Another won't form until all of a sudden, the wave changes to 2 [π]/2 within the same length, as the wave length of the buckle decreases and 2 buckles form within the same length, or once buckled, the lateral restraint can stiffen with further growth being forced axially, overcoming axial friction and the same buckle length gets longer, axial friction holds, lateral restraint gives and the pipeline pops up out of the ground.

Independent events are seldomly independent.