Hi to you all:

Maybe What I am asking is very simple, but I have no long experience with it, and after looking for it and asking some people I have not been able to find an answer; the question is the following:

I have an 8" 100 meters long steam pipe (250 ºC) and a hammer caused a displacement of more than 200 mm and all the shoes in this pipe on the rack are "out of place", I mean, displaced and trapped with the rack beams; so; they want to stop the line for sometime to fix it, and they want to know the behaviour of the pipe during the transition from 250ºC to 21ºC (ambient temperature), something like: "how long will it take until the shoes break (or the pipe)? Is there any way of making this type of transitory study with Caesar? If anyone coul guide me with this.

thank you very much.

Jorge Peso.

hello Jorge,

I would like to propose you to expand your normal input data by following:
1. add limit stops with gaps to the involved supports - considering pipe rack beam width, pipeshoe length and initial pipeshoe location.
2. add CNODE to that supports refering to displacement input for pipeshoe hight and axial movements.
3. add new load case(s) for the trapped situations (using displacement definition)
After that you should be able to discribe the piping before and after that accident for different temperatures.

If there was happened no damage of pipeshoe connection to steel structure but only the large unexpected displacements I recommend to check that piping system regarding missing static/dynamic anchor points / limit stops.

regards

I would say that if you have not seen the line by yourself before you start doing anything with CAESAR go to site and inspect the line. Assuming carbon steel line the expansion is about 3 mm per m. Check how the shoes are hooked, where the fixed points are, what type end connection and what happens to the line when it cools down. First set of tools: measuring tape, calculator and brains. Last set of tools: CAESAR, sorry COADE.

Last pictures of similar case I saw (about 3 weeks ago) showed beams bent and shoes collapsed. Straight pipe was fine.

Jouko:

I have already seen the line by myself, I know how the shoes are hooked and what I want to know is that exactly: how the line behaves when cooling down.

I have a clear idea on my mind of wht to do, which supports to remove, where to add fixed points, etc.... What I wanted is the way of making some cheking, some kind of transitory (time history) analysis of the line going from 250 ºC to ambient temperature, just to know if we can fix it the way we want to make it; I mean, just for knowing if the pipe will resist that cooling down by stresses having those shoes hooked.

Thanks.

Jorge Peso

You have the physical geometry and pipe data for each support ready for input.

Try running 6 (or 8) cases: One for cold, one for each 30-40 degree temperature change between cold and final (hot) condition, and one for the final hot condition.

Print the results, you have the intermediate conditions at each point. Seems that something else is off, some eqpt is moving at an end point: 200 mm total displacement seems to be lot for such a short total length and low temperature change.

Notice that if a hammer (impulse) impact jerked the ends of the pipe (enough to damage the supports and move the pipe out of the support shoes) then a gradual cooling checks won't show the effect of the water hammer.

It will show the steady, gradual movement that is expected, but not the impulse event.

We tried last year to do similar calculation. Too many unknowns and at the end line moved differently to the calculations. You must have somewhere change of line direction. When the line starts cooling down the weakest element will start deforming. This can be your expansion loop, rack beam bending, rack beam web deforming, beam foundations... Maximum force acting should be cross pipe metal area * yield * maybe 1.2. Minimum force = 0.

If you can figure out which points in the line are fixed then you can calculate few load cases with different metal temperatures (a bit lower than actual steam) and present the movement results in graphical form. You can model that 200 mm using CNODE and offset the node. Using movement data try to figure out which would give first/which is restricting cooling down. I would try to remove any item that is restricting as it will be more difficult to do after the line has cooled down.

If you try to calculate the max stress in the line and give figure lets say 150% the the first question customer asks is "What is the safety factor in the code?" I refused to give stress level.

There is only one thing that you can be sure of. Straight pipe sections will get shorter what ever is on their way.

That line is now like a loaded spring. Everybody should be careful when working on it.

One possible way to do the repair is to get the pressure off. While the line is still hot pull it a bit more to unhook the shoes (If you can), lift it up and let cool down. Customer did lifting of over 30" superheated steam line while on operation.

Ok thanks Robert;

The gradual cooling of the pipe will not affect so much.

Finally I think we are just gonna fix the supports and re-place the pipe on its place until the next jerks comes :P.

thanks.

Jorge Peso

I'll jump in now and give my 2 cents.

I think Jorge's idea regarding fixing the supports and replacing the pipe is the best idea. As Jouko stated above "the last tool is CAESAR II ". Actually I'm not convinced that you can use CAESAR II for this at all. Here is why ...

You had an event that jumped the pipe and broke supports. Now, perhaps you could run a number of dynamic events (or static approximations) to overload the pipe. But, CAESAR II will not break supports. To do that, you would have to change your guides to bi-linear restraints with 1 lb/in stiffness after the max load (break force) was attained. (This eliminates the use of the dynamics modules since restraints have to be linear.) After some number of approximation runs you have the pipe in the displaced position (maybe).

The next step is to turn the system off - temperature goes back to ambient and pressure goes to zero. Unfortunately this is really a different model, without the broken restraints and in an initially displaced / stressed position. This is called "load stepping" (where one load cases starts at the end position of a previous load case), and CAESAR II can't do this.

Perhaps you could invent some set of CNODES with displacements to approximate this behavior, but it would be messy. Everything goes out the window if the pipe yielded in breaking the supports in its initial movement.

So again, I think replacing the pipe is the safest thing to do. However, I would suggest a few "what if" type dynamic runs to see if you can determine why this happened, and make necessary adjustments to prevent this from repeating.

Well, thanks to everybody:

This time we have decided not to use any simulation program.
Getting cold the pipe or the supports are not going to suffer more than what they have suffered before. So we will lift the pipe, replace the broken supports, and re-locate the pipe on its position.

Anyway I have realised that the simulation with caesar II would not be real, for: this pipe is insulated and this supports hooked on the rack beams are lying on the beams but not the shoe or the pipe just the insulation lies on the beam, so in some part of the pipe the insulation deforms a lot in others it deforms less, that´s why we are getting strange results: like displacements quite different than the ones we can see in the pipe.

Thanks again to everybody.

Regards:

Jorge Peso.