Here's an example for context (none of these numbers are real scenario, just used for simplification):

A 4" line is routed with many bends in a rack on 3" wide steel columns at 25ft centers sitting on 12" long, welded pipe shoes. These spans result in deflection (0.65") in the pipe greater than what we would allow per internal company standards (0.5" for a 4" pipe). One philosophy I've heard over the years is the "bending" or "sagging" in the piping does not begin until after the pipe shoes, so modeling a pipe shoe as one single restraint is not accurate. Rather, according to this philosophy, it should be modeled as three +y restraints, spaced 6" apart to represent the entire length of the pipe shoe.

Now, per this philosophy, it is a 24' span despite the location of two of the +y restraints are not in contact with the steel and deflection is now 0.5" and deemed acceptable.

Is this philosophy an acceptable means of modeling? I can't wrap my head around this philosophy, please let me know if you agree or not.

Thank you!!

This is one way you could model shoes to get different effects. I'm not sure why you would need 3 and not just 2. I'd be more comfortable with them spaced at the width of the beam, and not the shoe.

Another way is to model the singular support as not allowing any rotation at the location, as well. If your vertical displacement is excessive one side but not both, this could also change the result.

But questions you should ask yourself:

• Why is sag important? Is it a safety concern?
o Is it to guard against pooling of fluids, ergo slugs?
o Is it to prevent pipe from running into another pipe?
o Is it for flow induced vibrations?
• Is the client cost adverse or risk adverse? More supports, or assume shoes add rigidity for fewer supports?

Edit to add: Your method assumes that the pipe shoe provides practically infinite rigidity to the pipe, and you may lowball some stresses. In addition, you'll be better able to back out moments, which may not normally come up, but these are generally minimal.

Personally that seems a bit dicey to me. I can understand where it makes a certain sense to use multiple supports, but with only 3" wide steel channels there would not really be as much stiffness and restraint at the two end points. (If you spaced them only 1 1/2" to either side of center, maybe...)

For further consideration are you modeling the steel that comprises the support? CAESAR II supports assume 1e12 psi for stiffness, so in reality you will actually have less. This will lead to potential deflection in your pipe supports.

I think modeling the support in more detail to support a desired outcome is not a conservative solution. If you start going into more detail like this, I think you should also model other details, which would likely cancel the benefits of the extra modeling in the first place.

Also Michael made some very good points about why you might want to be concerned about deflection.

Overall, only you can evaluate how critical it is if your assumptions make the line acceptable.

Thanks,
Duncan

I like the way this conversation is going as it points out the difference between modeling and "reality". We all must swallow many assumptions in initial design.
As a fellow piping engineer often states - we are not building watches!
Now, if I had an installation in the field where more specific data is available and specific questions to answer, I probably would start to tighten up my model to better correlate my digital model with the (analog) reality. In the case of a sensitive pipe shoe, if I saw it tilted on an edge, I would probably model the restraint at the edge rather than the center. If the load was large and the shoe sitting on some structural span, then the support stiffness may also come into play - for this system evaluation.

Thank you all for the input. I tend to agree, especially with Duncan's comment that you start to cancel out a conservative solution if you are modifying only certain parts the overall problem to meet a solution.

And for everyone's info, we have general guidelines (spans limits, guide spacing, sag limits, expansion limits) for piping design/analysis that if we are within these acceptable ranges we can almost nearly ensure ourselves of a quality design. We check sag for all the reasons Michael mentioned. If there's a particular line that economically doesn't make sense to try to go out of the way to meet all of these guidelines then we can single it out.

Dear Michael,

I agree to your words "Another way is to model the singular support as not allowing any rotation at the location".. You mean to say that torsional moment should be arrested ?

What I understand is, in large diameter pipes (>30"), Shoe width will be considerably high. So, it won't allow pipe to twist. Thus, We should arrest the torsional moment and the resulted moment should be given to civil for stiffer design of structural member and foundation.

Is my opinion goes along with your understanding? Any other opinion gents?

A typical vertical support... +Y with or without friction, effectively models a sliding pin connection on an infinitely thin beam.

While this is not physically accurate, shoe or otherwise, it's the "gold standard" on which the industry practice is made.

As such, you should feel empowered to look into it with a little more detail if and when you think it'll hurt you, but you could be setting yourself up for failure if you start doing so in order to take credit for it to reduce stresses and nozzle loads. Because once you go down that rabbit hole, you might be asked to figure out what happens now that you've calculated uneven loading effects on pipe shoes.

Either way, you should document that what you're doing is deviation from standard practice.

But to answer your question directly, Ankit, yes, there is a "toppling" resistance associated with the shoe that affects rotation in both directions. I suppose there is also be a stiffening effect to prevent pipe twist, as well.