Second-order responses of a FWT (using AQWA for the frequency domain analysis)

Wind & Water Computer-Aided Engineering Software Tools
1

I am working on comparing the experimental results from model tests of a FWT with the numerical results from OpenFAST. Because I am interested in looking at second-order responses and the necessary input files for the potential flow solution are not available in the literature, I had to develop my own hydrodynamic analysis. For the hydrodynamic frequency domain analysis I used AQWA and then translated into the WAMIT format using BEMRosetta.

I think my HydroDyn input file is okay and looks like this:

image
image892×747 39.8 KB

The error message I get is the following:

…"Calculating second order difference-frequency force using the full quadratic transfer function.

FAST_InitializeAll:HydroDyn_Init:WAMIT2_Init:DiffQTF_InitCalc: The second order WAMIT data in
.../HydroData/AQWA_model_FQTF/Analysis.12d is too sparse for the
interpolation routine used in the full Difference QTF calculation. At some later point, we will
allow for sparse data when a different interpolation routine is implemented.

OpenFAST encountered an error during module initialization.
Simulation error level: FATAL ERROR

Aborting OpenFAST. "…

I am wondering why this could be, what does it mean?

In my hydrodynamic model, I followed the suggestions and recommendations from section 4.2.8.4.8.4. in 4.2.8.4. Modeling Considerations — OpenFAST v3.5.1 documentation : The discretization of the set of directions is fully populated (not sparse), with a range of [-180:10:180]. Also, my frequencies span the range of [0.25:0.05:2.75] rad/s.

Thanks,
Tomas.

2

Dear @Tomas.Lopez.Olocco,

If I understand correctly, your 12d file is fully populated for every combination of frequencies in the range [0.25:0.05:2.75] rad/s and every combination of wave directions in the range [-180:10:180] deg for both wave components; can you confirm? Do you also have all 6 modes for each combination?

Best regards,

3

Dear @Jason.Jonkman,

Thanks for your quick reply, I wanted to take my time to be 100% sure of everything and take some tests.

Yes, as you say, My AQWA model is fully populated for every combination of wave directions and frequencies, and I have all 6 modes of motion.

image
image1271×557 69.2 KB

The .12d file looks something like this:

image
image1313×601 38.2 KB

Maybe in my translation process, I made a mistake but basically, I have something like this:

[T_m1, T_n1, \beta_m1, \beta_n1, DOF, QTF_mag, QTF_phase, QTF_real, QTF_imag;
T_m1, T_n1, \beta_m2, \beta_n2, DOF, QTF_mag, QTF_phase, QTF_real, QTF_imag;
… … … … … … … … …
T_m1, T_n1, \beta_mn, \beta_nn, DOF, QTF_mag, QTF_phase, QTF_real, QTF_imag;
T_m1, T_n2, \beta_m2, \beta_n2, DOF, QTF_mag, QTF_phase, QTF_real, QTF_imag;
T_m1, T_n2, \beta_m2, \beta_n2, DOF, QTF_mag, QTF_phase, QTF_real, QTF_imag;
… … … … … … … … …
T_m1, T_n2, \beta_m2, \beta_n2, DOF, QTF_mag, QTF_phase, QTF_real, QTF_imag;
… … … … … … … … …
T_mn, T_nn, \beta_mn, \beta_nn, DOF, QTF_mag, QTF_phase, QTF_real, QTF_imag ]

I’ve discovered that if I only keep the 0deg wave heading direction in my .12d file, OpenFAST can calculate the 2nd-order forces using FQTFs. Therefore, maybe the problem is there, in how the matrix is populated. How does HydroDyn populate the force matrix from .12d and .12s files?

I read this file (https://openfast.readthedocs.io/en/main/_downloads/ec7b205d06be5a23fd2d7de23c517805/HydroDyn_2ndOrderForces_Plan.pdf), but I might be missing something.

Thank you!

Best,
Tomas.

4

Dear @Tomas.Lopez.Olocco,

When you have multiple wave directions, do you all combinations of these directions to ensure that they are full populated? (I don’t see that in your example.) For example, if you have the wave directions of 0 and 10, you must have the following pairings: 0,0; 0,10; 10,0; and 10,10.

Best regards,

5

Dear @Jason.Jonkman

from your example, pairs (0,10) and (10,0) are missing. That must be it. I will take a look into my translation process. Thank you!

Best,
Tomas