G'day Everybody,
Currently modelling a gas line that starts from a wellhead with the following known displacements: At initial start up the wellhead rises by 5inch due to thermal growth and during operation it moves laterally by 3inch on either direction. Just a bit confused in chapter 3-3 of the application guide where it says that "Once a degree of freedom has been fixed in one displacement vector, it cannot be free in another displacement vector at the
same node (leaving a displacement field blank will default to zero in this case)." How I modelled this is I placed a 5" displacement on all of the vectors (ie Vector 1, vector 2....) as the 5" displacement will be true for all ope case. I then placed a 3" displacement on the Z direction for vector 1 then a -3" displacment for Z of Vector 2 while making all the other values 0. Is this correct? Thanking you all in advance.

The answer to your specific question "Is this correct?" is "yes, I believe so".

As to the statement "Once a degree of freedom has been fixed in one displacement vector, it cannot be free in another displacement vector at the same node (leaving a displacement field blank will default to zero in this case)", think of it this way:

Say you have a double acting "Y" (vertical) support that for some reason grows up by 0.5" in the operating condition. There are several points to recognize here:

1) Numerically speaking, a restraint is displacment with a magnitude of zero.

2) If you define a boundary condition (restraint or displacment) at a node point, that boundary condition exists regardless of the load case in question. (You don't run out to the field and add or remove a restraint before changing temperatures or pressures.)

So, if in Displacment Vector 1, you define DY=0.5, then you have fixed "DY" for all nine displacement vectors. If you don't specify a numeric value, CAESAR II will assume 0.0. You can't say that for this specific node, it has a displacement in one vector but is free in another vector. (That would mean you ran out to the field and changed the system.)

The same assumption is made when you analyze the Sustained Load Case. The load primatives making up the case would be W + P1, note the absence of D1. Because you defined "DY=0.5", but did not include "D1" in the load case, CAESAR II will assume 0.0 in the DY direction at this node point for the Sustained case. (Again because the boundary condition must be consistent.)

Hmmm... So why will I assign more than 1 displacement vector then? That is if a displacement is tru for all vector then will all displacement vector 1,2,3.... be all the same? Please clarify? Reason is that I have been given two observed displacements that moves towards the Z and -Z plane.

You use as many displacement sets as you need. Usually if you have one thermal analysis, you have one one displacement set if you are modeling a boundary condition that moves due to equipment growth. If you have two temperature sets, then probably two displacement sets. CAESAR II will recommend T1 & D1 together and T2 & D2 together. And so on...

If, however, you are modeling anchor settlement then you will save a displacement set just for that and maybe include it in all analyses.

In your case you may want to run something like W+P1+T1+D1 and W+P1+T1-D1 if the displacements are all opposite. Otherwise put the +Z's in D1 and the -Z's in D2 and use them where you want (e.g. W+P1+T1+D1 and W+P1+T1+D2).

Always check the results to confirm that your boundary conditions react as you expect.

Cheers guys!

I'm not entirely sure Richard has nailed the question here. If you define D1 = 0.5" in Y, all the other displacements will be set to zero at that node in each vector UNLESS YOU CHANGE THEM. If you delete the displacement fields, the node is free to move assuming no restraint at that node. Although defining one displacement vector also activates all eight other vectors at that node, you do not have to use them but you do need to remember that once the displacement field is ON, the node always has a defined displacement, whether it is zero or some other number. You cannot have free movement and defined displacements at the same node.

Use each vector to define an unique displacement and then you are free to combine them at will.
since it sounds like you're doing an offshore wellhead given the lateral displacements you are describing
One typical set of vectors [ignoring rotations] might be:-

X Y Z
D1 0 5 0 wellhead in OPE condition
D2 3 5 0 wellhead in OPE condition with wave action in X
D3 0 5 3 wellhead in OPE condition with wave action in Z

Or you could define
X Y Z
D1 0 5 0 wellhead in OPE condition
D2 3 0 0 wellhead wave action in X
D3 0 0 3 wellhead wave action in Z

and then combine them in the load cases
W+P1+T1+H+D1+D2 - OPE conditions with hangers and wave in X direction

Many guys just double the single lateral displacement [eg D3 = 6" in X or Z] in order to get the full stress range for fatigue calcs in one shot, but you do need to be sure that there are no non-linearities in the system.
This would define:-

X Y Z
D4 6 0 0 Full wave action range [simplified] in X
D5 0 0 6 Full wave action range [simplified] in Z

By spliting the cases up you can then calculate what the fatigue life for each one is and use some sort of cycle counting to combine them all together into a final fatigue life. The other thing to watch is how the conductors are guided. If a circular guide is used, the wellhead will only be able to move 3" in any direction. If 'X' and 'Z' chocks are used there is the possibility of the wellhead moving SQRT [x^2 + Y^2] = 4.24" in any 45degree direction. As a comment, 3inches of lateral movement does seem a lot though; The North Sea generally has much smaller movements [< 1/2"] after the conductors are chocked down.
It is also worth considering adding a dynamic amplification factor [2 to be conservative] to the horizontal movements as wave action is a dynamic action and not a static one.

May I ask another related question... just looking at the restraint summary report (while confirming that the reactions are as expected), under the maximum values, what does the /2.... or /3 or /4 stand for? Eg Max FX = 300/4. Are these sums divided over two, three? Why will this be the case?

Max FX = 300/4. => 300 in load case 4

CAPT, where can I input (or how do do I model) the dynamic amplification factor of a wave motion?

Ah! there lies the problem. Unless you have tons of time and tons of enviromental information to hand it is pretty much guess work. There are so many variables. Wave period is probably the biggest consideration as the wavelength [hence period] and height are related in the open sea, but there are always smaller waves superimposed on top of the bigger ones to confuse the issue and then you have the relfection and splashback from other risers and the jacket to consider.

Taking a 'dumb' DLF of two may seem a bit of a cop out, but the superposition of different modes can throw up some strange behaviour. I remember seeing one offshore wellhead that danced around vigourously for ages, but every now and again would 'go mad' for a few seconds. Whether it was slugs of water or sand coming up or some combination of modes coming together I don't know. I've never had the time, resources or mentors to hand to chase this thing through fully.

You could try to define a response spectra or similar, but it is something I have never done [or seen done]. Generally be a bit more cautious around isolated weights like drain valves or small bore [ie low stiffness] lines with valving, where there is little to stop them 'bouncing'and they are the 'tail being wagged by the dog'. The main flowline may not need a high DLF if the conductors are chocked down to reasonable gaps.
Personally we do the fatigue calcs outside of Caesar in here and we apply the DLF directly to the stresses being considered as part of the fatigue calc.

Conservatism in this area is good. The cyclic endurance required of these lines is mindblowing; ~1E9 cycles upwards with movements of +/- 3" is not small beer and the consequences of a failure from fatigue or otherwise does not bear thinking about. Nobody ever got fired for not making a platform go on fire.

Thanks CAPT. Rest assure that what I am analysing is not live and that the 3" lateral movement was only an example. I modelled it this way to see a more "pronounced" animation in Caesar. Two things that I would like clarified rather than assume... DLF and Conductor.

A staba as to what you mean by conductor.... the actual wellhead where the gas line terminates to? I agree with difference in movement of circular versus aquare chocks.

Pickles
I have a terrible habit of swapping terms for things occasionaly, so apologies if my posts have swapped terms inconsistantly. The DLF is Dynamic Load Factor - this is the same as the amplification factor I mentioned in a previous post and is just the ratio of system response under dynamic loading compared to static loading and can usually be regarded as being between 1 and 2. If you hit resonance however, the DLF can become [in theory] infinitely large [for an undamped system].
The conductor is the structure the riser passes through just before the xmas tree. The riser is normally 'chocked' here to limit lateral movement to whatever to deemed acceptable.