piping support redesign

[DrejzaVu]

I have a 12" line that originally had a 19.5" long 8"-dia pipe welding on the bottom for a support. A new spool piece is fabbed for installation, but with a 6"-dia pipe for the support. The team has asked me to bless this. Although I do not think there is any problem, I would be only be giving my gut instinct. How would I go about verifying it is acceptable? What calcs and considerations must I use? I have not done much in the way of piping supports like this.

Thanks.

 

[StressGuy]

Depending on the situation, you may have a pretty complex problem. The most common method for a manual review of the local stress is typically known as the Kellogg method (or some derivative of it). It's pretty conservative but is good for basic stuff.

If you've got no thermal movement to speak of and you're mostly looking at weight load, the 6" is likely to be adequate. You really need to look at the configuration to see how much weight the support point is carrying.

Now, if you do have thermal movement, you're going to be concerned with bending moments at the trunnion/pipe interface due to the frictional resistance of the trunnion as it tries to slide across the support point.

I'm not quite sure where to tell you to get a copy of the kellogg papers. Pretty much any decent engineering company is going to have a copy, or a derivative of it.

The other option is to run an FEA on the junction. FE/Pipe is a specialized FEA program that has templates to look at these kinds of stresses.

It's not something to be taken lightly.

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.

 

[DrejzaVu]

Thanks Ed,

Weight is the biggest concern as the fluid in the pipe is water at around ambient temperature.

The piping is only about 40" and has two 12" butterfly valves on either end, plus a large check valve. One end elbows down into the ground (then running underground). There is already an 8"-dia pipe for a rigid support about 10 feet down the line before connecting to a pump flange.

Would I be concerned about buckling of this 6" x 19" support pipe? It is not really a long, slender column. What failure mechanism am I really looking at and how do I interpret B31.3 to use the right calcs, if any?

Thanks again.

Mark

 

[KernOily]

Ed do you think he could get by with one of the analyses in Roark? It would be more conservative than FEA.

In my experience the Kellogg papers are drying up and blowing away, and those that haven't yet are falling off the flat edge of the world. I have yet to find a copy and man would I love one. I barely found a copy of the Markl papers. They are disappearing as the stress principals from the old companies retire and go fishing and as the companies continue to merge. Me, I have to get by with Roark and FEA. - Pete

Thanks!
Pete

 

[DrejzaVu]

I guess I can simply add up all the weight on top of the two supports and then divide by 2 and then divide by the cross-sectional area of the pipe to get the compression stress in the pipe. Then what would I compare that value too to see if it is too much? Are there compression strengths published for 12" CS pipe?

Mark

 

[DrejzaVu]

Ed and Pete,

Here is what I did. I looked up the chapter on buckling and columns. The theory is that if the load is applied along the axis of a short column and not eccentrically, then buckling doesn't occur. Instead the critical unit load is equal to the yield strength. So can total up the weight of all the piping, fittings, valves, and water in the pipe for the size/length of the pipeline and divided that weight by the cross-sectional area of the support pipe to get the stress in the support. If this stress is lower (which it is by a factor of over 10), then the support is strong enough. Does this sound like a reasonable calculation?

Mark

 

[StressGuy]

Column buckling is something worth checking in the support, but, particularly as short as they typically are, that's not the failure mode you are going to be concerned with.

The critical point is at the interface of the pipe and support - the circular projection where you are welding the two pipes together. The force due to the weight of pipe is going to cause bending in the main pipe. You can picture a piece of saran wrap, restraining it by the edges, and then pushing in the middle with your fist causing it to deflect and stretch out.

I'd have to go through my Roark book, but I don't know if he's going to have something that would essentially be a cylinder under internal pressure with a uniform load applied to the outside in a projected circle.

Since you are only looking at an axial load (i.e. along the centerline of your support), the Kellogg equation is this:

S = (1.75*f*sqrt(R*t))/(t^2) + P*R/t

f = W/(2*pi*r)

S = Bending Stress, psi
W = Axial Applied Load, lbs. (Weight in your case)
r = Radius of trunnion (support pipe), in.
pi = 3.14159
f = Load per inch

R = Radius of pipe, in.
t = corroded thickness of pipe, in.
P = Internal pipe pressure, psig

Your allowable stress is going to be 1.5*Sh, where Sh is your base code allowable stress at temperature. If you're B31.3 and using a common carbon steel like A106B or A53 B, Sh for temperatures less than 400F is 20,000 psi

Based on your description of the system, I don't expect that you have any significant circumferential or longitudinal bending loads.

I'm afraid my copy of the Kellogg paper is probably a 4th generation copy that is already in pretty poor shape. Otherwise, I'd see about scanning it and trying to post it somewhere. Perhaps one of our other distinguished members has a cleaner copy. I would expect it can be found in the sadly out of print Kellogg Design of Piping Systems. Unfortunately, I've not been able to obtain a copy of that book for my collection yet.

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.

 

[stanier]

From memory NozzlePro from this website allowed you to calculate this type of support. There was a download freebie that you could use but it was limited. In any event the spftware looks eminently useful for this and other applications.

http://www.paulin.com/prg/paper/NP06-2004.pdf

 

[StressGuy]

Nozzle Pro will work if you have a support on straight pipe. You just need to go to the advanced options and check "do not cut hole in pipe".

Paulin promised a Stanchion Pro program years ago to give a Nozzle Pro style approach to dummy leg analysis. As it is, you have to dig through the full FE/PIPE interface to due dummy leg checking.

Unfortunately, that one seems to have died on the vine. I think my company is probably uncommon in that our pipe stress group is the bigger user of FE/Pipe and Nozzle Pro as we due most of our own nozzle load checks ourselves instead of submitting loads to a vessel engineer. Paulin seems to target the vessel engineers instead of the pipe stress engineers where the programs would be more useful.

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.