Repository navigation
The dynamic response of the OC4 jacket differs significantly between OpenFAST 5.0.0 and Mechanical APDL 2025R2. #3490
Replies: 1 comment 3 replies
|
Dear @laitiandeng-maker, Taking the tower-base loads from OpenFAST as input to Mechanical APDL for a jacket analysis is a one-way coupled approach that is not common in the offshore wind industry due its potential for numerical instabilities and inaccuracies. More common approaches are the fully coupled approachavailable in standalone OpenFAST (with SubDyn/HydroDyn enabled), or the sequentially coupled approach, where a superelement from an FEA model is embedded within the aero-elastic tool, i.e., through ExtPtfm in OpenFAST. See a similar discussion in the following topic on our forum: https://forums.nlr.gov/t/jacket-analysis-using-fast-and-abaqus/1558. Given the potential for numerical instabilities and inaccuracies with a one-way coupled approach, I'm not surprised you are seeing differences between that and the fully coupled approach from OpenFAST. But is hard to know exactly what is going wrong with such a complicated model set-up. If you want to pursue this further, I would first suggest starting with a simpler case, e.g., no turbine on top, e.g., a lumped mass without aerodynamic loads applied at the transition piece (tower base); how do the two solutions look for that case? Best regards, |


Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Hi everyone,









I’m comparing the dynamic response at four nodes of the OC4 jacket between OpenFAST 5.0.0 and Mechanical APDL 2025 R2. The four nodes are:(-4.385, 4.385, 4.378), (-4.592, 0, -1.958), (-4.82, 4.82, -8.922) and (-5.967, 5.967, -44.001).
The compared quantities are velocity, acceleration, displacement, axial force, shear force, and bending moment.
The modal properties of the OC4 jacket are very close between the two models (see Fig. 1). I kept the linear-wave parameters in APDL consistent with those in OpenFAST (see Fig. 2; the two linear waves almost coincide over 45 s).
I extracted the tower-base loads at (0, 0, 20.15) from the .out file generated by OpenFAST:TwrBsFxt, TwrBsFyt, TwrBsFzt, TwrBsMxt, TwrBsMyt, TwrBsMxzt.
Unit conversion was already applied when importing them into APDL. I used the first 45 s of these loads as the load input for the OC4 jacket in APDL, applied at (0, 0, 18.15). After running APDL for the same 45 s, I extracted the generated CSV data and compared it with the corresponding data in the .out file (see the uploaded figure).
The curves show very poor agreement. I’ve been trying to fix this for a long time, but I still don’t understand why this is happening. Could anyone help me figure out what might be wrong? Has anyone run into something similar? Any help would be greatly appreciated. Thanks!
Bset regards!
All reactions