[JM92]

Dear all,

I'm currently trying to set up a frequency domain model in Matlab to calculate the RAOs of floating structures based on the complex amplitude:


Input values (added mass, damping, excitation force, stiffness, mass) are all taken from AQWA using NREL's BEMIO tool.

So far so good, and my first "verification" test was to recalculate the RAOs from AQWA. But the result of this confuse me completely:

• My first check was to calculate the heave RAO of a ship using only the values on the diagonal of the matrices (so no coupling-influence). This works well and matches the AQWA RAO perfectly.
• My second check was doing the same thing with the pitch RAO, which showed a small difference to the AQWA results (2-3 %).
• Both checks worked using scalars (using the values from the diagonal of the matrix) for 1 DOF each, and matrices (with all of the coupling values equal to 0) for all 6 DOF simultaneously. Both "methods" resulted in the same values.
• Afterwards, I used the full 6x6-matrices with coupling terms, and the results now differ strongly from the RAOs calculated in AQWA. For heave it's 5 % off, for pitch its 20 % off, which I do not understand at all.

I checked nearly everything I could imagine: Errors in input data (no difference between AQWA, AQWA database, matlab), inaccuracy in computations (16 digits are used in matlab, re-check with Python numpy, same results), my implementations being wrong (should no be the case as the 6x6 case without coupling works well), units (degree vs. radian, seems also fine).
After all, I'm writing this post, because I don't have any other ideas: Did anybody ever try a similar thing and encounter similar problems? Is there maybe a hidden "tweak" in AQWA to calculate the RAOs, which I'm missing by just using the equation I gave above?


Kind regards,
JM92