I'm looking at a couple of stress reports that deal with the fatigue life of offshore Xmas tree flowlines. These flowlines typically have to be very flexible in order to survive an extended period offshore, and are thus usually hung from constant effort supports. The riser from the well passes through a 'conductor' which limits it's lateral movement caused by wave action before connection to the Xmas tree. Therefore a typical spec would be +/-25mm laterally [and sometimes +25mm vertically] caused by wave action and +100mm upwards caused by thermal growth of the riser itself. The wave action is very frequent [say 10 waves an minute, for 20 years = 105120000 cycles], but the thermal growth, although much larger, may only be 1 cycle evry month or so.
The reports I have seem have taken the stress range caused by the movement from full +ve to full -ve wave action in each direction from a static analysis and used that to compute the fatigue life [how that is done is of course is another matter].
For those of you who have never been offshore, imagine a several tonne block of steel with piping coming off it [the xmas tree], dancing around as if it was being shaken by a giant and you'll get the picture of what goes on.

If we assume that a full dynamic analysis is not nessecary, and that we cannot be sure that every wave WILL NOT cause full displacement, do we need to apply a DLF to the stress ranges caused by these lateral wave movements when they are considered in the fatigue calculation? Personally I think that it is not required for the main flowline, but may well be good practice on associated smaller lines, such as drains, vent valves and injection lines that may well get excited by the main flowline as it shakes during operation [dog shaking the tail, as JB would say]. The addition of a DLF to these branches may well be a case of conservatively increasing the stress ranges being considered and hence reduce the fatigue life of certain parts of the system, but is that nessecarily a bad thing?

When they bury you inthe end if the worst thing said about you is "He was a sound conservative engineer" by my take I would say you have done all right.

DLF is influenced by period you might be able to cut down the DLF by looking at some natural frequencies. But yes ther3 is a DLF as the Dog wags its multiple small tails...

Take a read on this for some thinking....http://www.coade.com/newsletters/dec98.pdf

Note that the high cycle displacement condition is a candidate for using the f factor for an unlimited number of cycles that was added to the 2004 edition of B31.3, f=0.15. If the stress range for those conditions is less than SA with f=0.15, they are not included in calculating the equivalent number of cycles.

I would think that you would design your branch connections etc such that their natural frequency would be significantly higher than the 1/3 hertz excitation frequency (20 waves per minute) in which case the DLF would be essentially one.

Captain Kenny,

I have a similar situation. While I can't help with you query I was hoping I could ask you something. What is the best way to model the Christmas Tree/associated pipework. I have been told that I have +/-12mm lateral movement at the base of the Tree. The tree is approx 3000mm tall with production pipework at 1500mm from base and other pipework at 2500mm from base. Any advice would be appreciated.

JOD

We just model a rigid to represent the wing valve and apply displacements there. Make sure that all the displacements are filled in with the 'unused' degrees of freedom in any case set to 0. Essentially we are assuming that the tree does not deflect or rotate as it moves. Simplified to the bone, but getting relative stiffnesses of things like xmas trees is never going to be easy.
For what it is worth, let me know if you want any other input. I've done a fair bit on this recently and am happy to help if I can.

Thanks Kenny,

We anchored at the wing valve/pipe interface and applied the Displacements at a CNode. How/Where did you restrain the wing valve?

These things are short on space anyway but we are really struggling to introduce flexibility to cope with this movement. Any suggestions would be appreciated.

JOD

Answer, is it relevant to use DLF?
The simple answer is no. The wave cycles applied to any flowline or related lines is not just the theoretical wave cycle for any one year. The wave cycles are recorded for no of years and the worst case is summed for different wave heights. You can notice the maximum no of wave occurs at 0 - 1 meter level. However the applied cycle which is calculated as the summation of different heights at Tree terminal points, which itself is conservative approach to evaluate fatigue of the line. In my opinion the DLF is not required.