Virtual anchor length in my calculation is almost 140m as CII calculates it automatically in the buried conversion window but, in the CII output and when I check the Restraints displacements, the "virtually" fixed point seems to be only 40m far from the last tee connection. (I cut off the pipe line up to the calculated CII VAL)I know for a fact that if one models the pipe line up to 3 times of VAL, they could see the virtual fixed point almost at the length of VAL calculated by CII ( and indeed it happens in my calculation) what I do not understand is why my pipe line stops moving in shorter length? can anyone explain?
(I am positive that I have broken down the original elements to the proper lengths in case you may want to doubt that)

How did you determine that this point, 40m from a modeled tee, is the (beginning) of an inert section of pipe? Is there a single node that has zero movement or is this node the first of several nodes that do not move?

If you extend your model, does this "virtually" fixed point move? If not, I think your model is fine. If it does move, my guess is you need to extend the model.

the node in the proximity of 40m from the last tee has nearly zero movement; it is where the axial movements signs of the consecutive nodes are changing so I suppose the virtual fixed point must be somewhere in between my nodes...
as i said before, I extended my model to 3 times of the VAL calculated by CII and the virtual fixed point in my calculation fell somewhere in the vicinity of VAL.
my question is : is it like a rule of thumb to extend the model about 3 times VAL to have accurate results? or shall we simply cut off the model at VAL and put an anchor there?

thanks in advance

Rule number one. Do not automatically place an anchor at the place where you figure the pipe will stop moving axially.
The idea is to model enough straight pipe so that the analysis will be able to accumulate enough friction to stop itself. The VAL is an estimate of the length of pipe that can do this. With this straight length in the line, the upstream and downstream portions are structurally detached for analyses at or below the temperature/pressure for which the VAL was calculated.
Now, the VAL calculation in CAESAR II assumes a stick/slip response to axial friction - rigid axial restraint or sliding with a resisting load. But your CAESAR II soil model probably has a some sort of deflection associated with the estimated ultimate axial load - that will give a nonrigid stiffness to our "stick" in the stick/slip model. Therefore, some movement must be absorbed to build up the friction load that is accumulated down the line. For this reason - I suggest you run at least 2 VALs down the line before truncating the system.
I don't really care if it's 2 or 3 VALs, if I have that straight length in my model, it's just one pipe that you build and CAESAR II will take it from there. The important thing here, if you have less than 2 VAL between your last modeled thrust point (e.g., a change in direction or branch) and the next thrust point that you are trying not to model, I would code through that next thrust point.
I think 2 VAL should be enough but I have not looked at enough buried models to practically reinforce this concept.
Back to my inert section of pipe mentioned in the previous reply... I am not looking at a single point on the line that does not move axially; you will find a single such point in almost any run of hot pipe - buried or unburied. I am looking for a series of nodes that do no move axially. Now, in your buried pipe, our "Zone 3" (friction controlled) automatic spacing between added restraints is 100ODs. They cover a lot of pipe distance. And zero displacement, in my opinion, need not be 0.0000, If you get my implication.

thanks Dave

I would like to get a little more information on this. I have been running models in B31.8 on pig traps. Essentially a small section of pipe that has some 45 deg elbows coming above ground from the mainline to a barrel and then a tee going back below ground to tie back into the mainline. Nothing really to it.

But my problem is this. I have been modeling it with the VAL in mind. I restrain my belowground pipe and leave the above ground pipe unrestrained. We only deal with temps between -20 to 130. Pressures are generally around 850-1050 MAOP and hydro cases being a little higher. Mainline pipe sizes range from 6" to 30".

Some of my pig traps fail on the tee and some do not. Usually they have very little difference. I do use the soil modeler and I have been using real data from geotechnical analysis.

One of the main concerns that one of our Sr. Mechanical Engineers has here is that he feels that the extended length of below ground pipe is contributing to more stress than necessary on the tee. The reason why we say this is that in the real world, we've looked at a lot of the traps and do not see any signs of significant movement at the ground interface. Or stress cracks on the tees. We've had a lot of these in service for many years (10-20+).

I feel like I am following all the appropriate processes from Caesar to design these. And generally we do not have the option to extend pipe or add ells. There are lots of these traps that go in tight places. I have tried increasing the wall thickness and many many different restraint situations. As a large company that has been in the industry for over 50+ years, we do not use hangers/springs for this type of design. I'm not saying we shouldn't, but it's a fairly new idea for us to use hangers and would change a lot of practices. Not to mention, we already have many of these traps installed along our lines and have not seen any issues.

I am just curious what others might think and if there are any suggestions.

As you travel upstream of your pig trap do you have a very long straight run? Or do you have a change in direction soon after you run underground?
I ask because the idea of running a long run of straight pipe is to have enough straight pipe to fully develop (by accumulating friction) the necessary axial resistance to work against the thermal and pressure loads in your run. (that's what the VAL is all about.)
If you have a bend closer to the pig trap than your VAL, friction will not be your only resistance to your (T+P) loads. Instead, the pipe "around the corner" will bear against the pipe. That bearing is not necessarily rigid in term of stiffness. So instead of building up the total axial load (caused by T+P), the line may push the soil away (around the corner) thus reducing the residual axial load on your tee. Your axial load on your trap may be more controlled by soil bearing rather than (axial) soil friction.
Regarding field experience versus model results - I think the best path forward would be to try and get the model to match your experience rather than rethink your experience to justify the model.
Pipeliners don't know much about springs; with low thermal strain and little vertical piping, they have little reason to.

I apologize it took me a while to get back. I didn't realize anyone had responded. I probably need to check settings here to get a notification or something.

Anyway, typically we have long straight runs. There may be a small PI or something within 100 ft or so upstream. Typically I model the system with a long (500 ft or more) of straight pipe.

I have started testing this by placing anchor's in various places and checking the loads on the anchors and checking the change in the deflection after and prior to adding the anchor. I have definitely noticed that with the B31.8 code (and natural gas), the systems seem very sensitive to deflection. I.e. Prior to an anchor, the failing tee will have only .5" of deflection, where as after the anchor is added, it will only have .25" of deflection and pass the stress code.

I went through various stages to get a tee to pass stress. For example, I had a model of a 16" diameter pig trap that would not pass code no matter how many hangers, guides, restraints I added. It was failing on the main tee that branches off the mainline to the pig trap. It failed at almost 700% give or take. I tried increasing the tee wall, I tried changing the lengths of the pipe coming and going from the tee. It was obvious that the thermal growth was causing these high stresses. I had a VAL at about 250 ft. Still failed. I am pretty diligent in putting in the soil variables and checking that nothing is "weird" with the model (i.e. breaks that might have happened from the CAD transfer, rigid weights, checking and heading warnings, I have received a lot of help from a PDS consultant on building really excellent load cases, which includes modeling the impulse/slug from the pig itself, etc.). So after much discussion between me and the PDS consultant, we decided that placing anchors would be acceptable. It almost seemed there was no other way.

Dear lmkaminski82,

Sometimes cohesionless/loose soil may help you to prevent over-stressed barred tee near pig trap. Keep in mind that, pipe always pushes adjacent soil. These type of soil needs special considerations and construction procedures. You should also check the allowable loads on insulating joint before UG, if any.
You can also let the pig trap move a little bit in pipe direction. For this, you should also model pig trap and other connected piping.