Inconsistent RNA inertia

[Antoine-HENRY-setec]

Thank you very much for the package welib you developed. I was trying to use the example welib/fast/examples/Example_ExtractInertia.py and obtained several results that I can't explain. I've been trying to solve it without success so far.

1. Test on the IEA15MW Monopile
   I started from the IEA15MW Monopile model as downloaded from the official repo without any modifications. My goal was to obtain the RNA inertia matrix. On a first run, I got the following result:
   "
   RNA Mass: 947769.2264912041
   RNA Center of mass position (from tower top): [ -7.4485 0. 133.9985]
   RNA Inertia at center of mass:
   [[ 3.5174e+08 0.0000e+00 -1.4943e+07]
   [ 0.0000e+00 1.9265e+08 0.0000e+00]
   [-1.4943e+07 0.0000e+00 3.0948e+08]]
   RNA Inertia at tower top:
   [[3.7190e+08 0.0000e+00 1.7618e+07]
   [0.0000e+00 2.6540e+08 0.0000e+00]
   [1.7619e+07 0.0000e+00 3.6207e+08]]
   "
   At first I thought that the CMz = 134m was actually from platform or MSL and the string "from tower top" was incorrect. But looking into the methods, I realized that it was really supposed to be from tower top and therefore the value was incorrect. Investigating, I noticed that at some point a matrix Minv (extractPtfmInertiaFromLinFile) was extracted and then inversed. But the values found in Minv are very small (order of magnitude of 1e-15) which could easily lead to imprecise values in the inverse matrix.

2. Test on the IEA15MW UMaine Semi
   I assumed that the very small values obtained in Minv were because the system was too constrained to move enough in a the linearized simulation? So I tested with the floating model hoping it would improve. But I obtained the following values which I think are still irrelevant. I also noticed that apart from the CM coordinates, the inertia values had also changed by a lot, although it is the same turbine.
   ------------------------- RNA -------------------------
   RNA Mass: 18785761.022531055
   RNA Center of mass position (from tower top): [-0.3715 0. -6.1564]
   RNA Inertia at center of mass:
   [[3.6887e+10 0.0000e+00 1.0799e+09]
   [0.0000e+00 3.6777e+10 0.0000e+00]
   [1.0799e+09 0.0000e+00 2.3926e+10]]
   RNA Inertia at tower top:
   [[4.6263e+11 0.0000e+00 2.9196e+07]
   [0.0000e+00 4.6252e+11 0.0000e+00]
   [2.9196e+07 0.0000e+00 2.3929e+10]]

Could you please let me know in case you identify a problem in the way I use this method?
Thank you very much in advance

[ebranlard]

Dear Antoine,

Sorry for getting back to you so late. I would need to dive deeper to see what could be wrong. I'm pretty sure this used to work in the past. There could indeed be some small typos in the strings printed to screen. Linearization outputs are sensitive to OutFmt, but I believe the script takes care of that.
Let me know if you figure it out,

Sorry about that!

Emmanuel

[Antoine-HENRY-setec]

Hello Emmanuel,
Same here: quite a long time before going back to this subject! I've just looked into it again and I'm now fairly convinced of the following: the inertia are indeed correct and really given about CoG for J_G and about tower top for J_TT.

However, the CM given is not from tower top as written in the description of the function "extractRNAInertia", but from tower base. It can be inferred from the code in the following line:
TT_2_CM = np.array([CM[0], CM[1], CM[2]-PtfmRef2TwrTop])
Which gives correct numerical results.

So I think a quick fix would be either to "announce" CM to be provided from tower base, or to output TT_2_CM instead of CM.

Thank you very much for your answer in any case,
Antoine