Hi,

I am trying to develop an as-built model where pipe has displaced. I am trying to check the integrity of the piping in the displaced condition. I am attaching the hand sketch of how the layout and support is as built.

For instance, in the sketch lets consider S3, S4 and S5 has displaced towards east. Which means my pipe shoe has fallen from the base structural steel.

Case 1:- I know how much my pipe has displaced in vertical and horizontal directions. Should I model pipe support with Cnode and enter the displacements. In this case if the node displacement is 150mm, CNODE with 150mm disp input is considering a restraint and giving me restraint report. This is actually not true as there is no restraint at that particular location (pipe has displaced and pipe shoe has fallen from steel).

Case 2:- I know the restraints are not active and I am removing the support from all the locations where pipe supports are not in place with steel. We all know this scenario is also not true, as I am removing the restraints it leaves me with large unsupported pipe span. Hence more sustained stresses and Sags. I am confident case 2 is not possible. Is my case 1 more realistic approach?

I have more questions added to this model. Since my pipe has displaced it means the displacement is present in sustained condition itself. Should I have to build my sus load case with D1. In such case if I am building W+P+D for sustained and W+P+T+D for op case, when I am calculating exp stresses the effect of imposed displacement is not being calculated and stresses are in safe zone. But the other way round, If i consider D either in sustained or in ope case, the expansion stresses are above limit. I want to understand how to build a more realistic model. Kindly help me.

At the risk of sounding snarky, I would suggest that the use of CAESAR here would be better put at redesign and replacement of the line rather than the acceptability of a line that has already failed.

But as an academic exercise...
What are your displacements based on?
Is it sitting there, operating, hot, and they measured those displacements?
Did they shut off the line and then measure it?

The "right" model would be the one that would look like what would happen if you turned off the pipe right now and let it cool - with all its ugly effects and calculated forces, and then ran the model to happen to "arrive" to the present displacements if you turned the pipe on.

Edit to add:
A number of supports that have run off would look something like a +Y with gap, but for the shoes that might act as directional stops - only one for each linear run of pipe will actually function as such - and that'll be a substantial force in the sustained case - and it might have already permanently moved entire sections of lines, causing more problems and more slips.

So I doubt this will be a clear-cut answer since you have to perform forensics on what happened once the first shoe failed by slipping off, and this could have happened over more than one cycle.

Michael has put you on the right path. The answer is very much going to depend on where your system is now and how it got there.

We need more information to offer anything useful.

What caused the displacement? Your description of moved supports and sketch suggest some kind of hammer event.

You indicate that S4 displaced East, but also that is has an axial stop. Did the stop get torn off the support?

Is the system in operation, or has it been shut down?

The shoes that fell off the supports - are they pulling back and acting as anchors, resisting the pipes attempt to return to its neutral position?

You note that the fallen off supports are "not active" - does that mean that the pipe is floating above the steel where the shoes were sitting and not making contact now?

Has the support steel been distorted? It's pretty common in the cases that I've seen of a shoe falling off due to hammering that it bends the steel when it tries to snap back into its original position.

Hi Ed and Mich,

Thank you for your reply. Since I am working for operator company, people don't like the idea of replacing the pipe and it is also not an easy process. It has to go through a very long process and approvals.

The pipe supports got displaced because of some sort of slug event. Some rapid dumping to a drum via this header happened. During dumping the line faced slug forces which has caused the pipe to displace/move away from its supports. Since the pipe is not a continuous op line, the pipe was in displaced state even when it was not operating. This doesn't mean that the system has deformed.

I am curious on how to simulate the model if the pipe is bent or displaced.
Lets assume the piece of pipe is bent because of slug load for 50mm, which means pipe support is offset from bottom of pipe by 50mm. Should I model this as pipe with restraint and a cnode with 50mm displacement. Or simply remove restraint and enter 50mm displacement or simply remove the restrain and consider the pipe as unsupported. If I consider pipe with Cnode and enter the displacement, That is not true as there is no restraint at that point. At the same time if simply remove the support, the model will consider the pipe is straight which is actually not true. I am kind of trying to figure out a more realistic simulation.

Hi,

How about doing a reverse engineering. You assume a slug force and apply it , then check the displacement. From there, you can scale up or down the slug force until it matches the displacement that is observe in the field. Then you can check the stress developed in your piping system with this slug force.

Just my idea.

Cheers!!!

I have experimented this for another similar issue. I tried the first model with site observed displacements. Later I crated a duplicate model and carried out different iterations with various force values to observe similar displacement pattern, but it was very difficult and even after 10-12 iterations site observations couldn't be captured. But that is a nice thought. Thank you.

Hello. Maybe there is already excessive deformation in the pipe (plastic state) and behaving non-linearly and at this state it's have to capture using a beam element software. Just assuming.

If it's caused by primary based load (then there is a tendency not to redistribute stress). But elastic-plastic analysis using FEA may capture the effect after yield, and since the piping is still intact, you would expect the analysis to converge, but low cycle fatigue will complicate if present in the system (if the system will experience it over and over again). This can be a good candidate for fitness for service analysis which I haven't tried before (determining how many cycles the system will endure).

Warm Regards,