I have some piping on an offshore platform. I have data consisting of a maximum rotation, lateral acceleration, and vertical acceleration for the roll and pitch cases. I'm comfortable with how to combine cases and evaluate fatigue, but what I'm confused about is how to enter the motion. I am torn between
1) Putting the rotations in as displacement vectors at a cnode on each restraint, or
2) Putting the accelerations in as a uniform load.

Second option. The values will be given to you by naval engineer. They can convert the vessel motions into equivalent accelerations.


Regards,

I had rotations and G's. Tried #1. I put a Cnode at each restraint and fed in the rotations given. Got very strange results like unrealistically large displacements in the sustained case, even though it didn't have the D1, D2,... loads.

I'm going back to try option 2. Wish me luck.

Remember to insert zeroes for the other translations or rotations.

Regards,

Thanks! Now the sustained and operating cases are behaving correctly and passing. The pitch/roll/wind cases are failing by thousands of percent. I am going to use the g-force method for comparison. The Caesar manual specifically calls for putting in displacements for this type of load, but I'm wondering if I'm not implementing it correctly. Should all of the restraints be rigidly connected to one another and just have one single anchor point rotating, perhaps at the center-of-gravity of the platform? The way I have it now with individual c-nodes on each restraint may be the problem. If each restraint rotates about itself then the differential movement would probably cause the tremendous stresses I'm seeing.

No. That's not right either, because each restraint now deflects in the sustained case as if it were actually supported by a long pipe from the origin with 10x the wall thickness rather than the much-stiffer steel frame that actuallys supports it. I need to resolve the rotation of each restraint point about the CG into Cartesian displacements. Spreadsheet time!

Try g-force.

Regards,

I should have listened to you in the first place, but I was trying to follow the method in the Caesar manual. I put in the displacements on all the C-nodes and ran the analysis. The motion looks right - the system looks just like a piping system swaying on a floating platform. The stresses are enormous, however, and I am not convinced that they are correctly calculated. I don't think Caesar's displacement function is really intended for movements like this, but rather for slow movements (maybe quasi-static is the word) that are similar to thermal expansion or static deck deflection from the weight of equipment. I'm not sure putting g's in as uniform loads is quite right either, but it seems more realistic, in the vein of doing static seismic analysis. Is the real answer to do some sort of dynamic analysis? I don't have time to do any more experimenting at this time, so right now I'll throw some accelerations in. Oh well, I learned something by trying it.

I just wanted to close the loop on this thread by saying that putting pitch and roll accelerations in as uniform loads seems to work quite well, while the displacement-based method in the Caesar manual does not. Thanks to DanB for his accurate advice.

Welcome.

Regards,