Dear Jonkman,
I fixed the issue by rolling back the Fast.farm version to 3.4.1.
Thank you!
Sincerely,
Dong Zhiyan
Dear Dr. Jonkman,
As you advised, I tried to output wake slices of wind turbines by enabling OutAllPlanes.
In spite of activating yaw control of upstream wind turbine, the curled shape of velocity fields were not observed.
However, in low-resolution box visualization, the deflected wake of WT1 is observed.
Is it helpful for the issue of high-resolution box visualization?
For the detail, I attached wake slices of two turbines where the WT1 is yawed 30 degrees and low-resolution box.
Sincerely,
Wonsuk Han.
Dear @Wonsuk.Han,
Which wake plane are you showing for T1–is it the one closest to the rotor? I would suggest setting k_VortexDecay to 0.0 to see if that helps.
FYI: I still see signs of a numerical instability in the wake calculation of T2.
Best regards,
Dear all,
Can anyone share the N3 inflow wind generated by sowfa carried by FAST.Farm at the beginning?
Best Regards，
Dear @Jundong.Wang,
I believe you are referring to the FAST.Farm model with SOWFA-generated inflow available through my Google Drive: Neutral_8mps_3WT_Short.zip - Google Drive.
Best regards,
Dear Dr. Jonkman and members,
I am using FAST.Farm with version of 3.4.1 to calculate the load of a wind farm. The wind speed is 5 m/s and the turbulence intensity is 15%. However, I encountered the following problem in my calculation.
`Time: 171 of 800 seconds. Estimated final completion at 15:12:16.
T6:FARM_UpdateStates:WD_UpdateStates:ThomasAlgorithm:Tridiagonal matrix is not diagonally
dominant, i.e., abs(b(i)) <= ( abs(a(i))+abs(c(i)) ). Try reducing the FAST.Farm timestep.
Aborting FAST.Farm.`
The Turbsim .inp file and .fstf file are shown in the following figure.
I am looking forward to your reply.
Thank you.
Best regards,
Dear @Yuxuan.Chen,
Are you seeing anything unexpected in the wind turbine or wake response up until this error is generated?
I suggest trying a few things:
- Upgrading from FAST.Farm within OpenFAST 3.4.1 to the newest version of the software, currently within OpenFAST v3.5.3.
- Reducing the time step (
DT_Low) as suggested in the error. - Reducing the cutoff (corner) frequency of the low pass filter (
f_c) in FAST.Farm.
Best regards,
Dear @Jason.Jonkman,
I reduced the cutoff (corner) frequency of the low pass filter (f_c) in Fast.Farm. This time Fast.Farm terminated normally. Thank you for sparing your valuable time to answer my confusion.
Best regards,
1 Like
Dear Dr. Jonkman,
I could see curled shape velocity field in wake slice setting k_VortexDecay to 0.0.
It was set to 0.1 to follow your work’ A time-varying formulation of the curled wake model within the FAST.Farm framework’ in Wind Energy. I still do not know why curled wake is not shown in the 0.1 setting.
And I have one more question. Now I am working on dynamic yaw control by using UserYawCont with sinusoidal move. In low-resolution vtk, the wake meandering looks unstable to me. Is it derived from the numerical instability in wake calculation? (Still no wake generation in high-resolution vtk files.)
Best regards,
Wonsuk Han.
Dear @Wonsuk.Han,
Regarding k_VortexDecay, the cross-flow velocities used in the curled wake model will dissipate with downstream distance when k_VortexDecay > 0, and the higher the value, the faster the curled wake shape will dissipate.
Regarding the odd visualization of your wake, I would guess this is related to the relationship between the cut-off frequency (f_c) you selected and the frequency at which your yaw angle is varying sinusoidally; f_c could filter out the sinusoidal motion if it is set too low. Also, the low-resolution time step (DT_Low) impacts how resolved or coarse the wake is computed.
Best regards,
1 Like