Hello,
I am trying to get familiar with OpenFAST-HydroDyn, more specifically, the stand-alone HydroDyn driver. The version I am working with is OpenFAST-v3.5.3, the precompiled binaries. The model I am working with is the OC4-DeepCwindSemi.
I have two questions:
- When running the ‘potential flow only’ i.e. by not assigning any members/joints, the output motion remains zero. The supplied WAMIT files are used. The curious thing is that there is a net HydroFxi, but PRPSurge and B1Surge remain zero. See the snippet of the output file below. I also added a snippet of the driver input file and the OC4 semi-sub model. In the latter, the elements from ‘2nd order floating platform forces’ are not shown, but also not used.
Why does the solution not show any movement despite net forces are present? Must it be coupled to ElastoDyn?
- How to retrieve the loads, and eventually stresses, in the semi-submersible structure? For example, consider the lumped mass model of the OC4 below. Is it possible to compute the loads using the potential flow model on each of the elements. Eventually, the idea is to feed these forces into a dedicated FEM solver.
Dear @Luc.vanBeek,
Regarding (1), the standalone HydroDyn driver does not compute motion as a result of hydrodynamic loading. Rather, in the standalone HydroDyn driver, the motion is prescribed and the hydrodynamic loading including the effect of the presecribed motion is computed. In your case, though, you’ve set to use zero motion (PRPInputsMod = 0). To compute motion as a result of hydrodynamic loading requires the use of HydroDyn coupled within OpenFAST.
Regarding (2), the standalone HydroDyn driver does not compute hydro-elastics; rather, the substructure is assumed to remain rigid. As such, you can’t use the standalone HydroDyn driver to compute internal member-level loads (or stress/strain). HydroDyn does allow you to distributed the hydrodynamic load across the body–either through the strip-theory solution, or in the potential-flow solution, but using more than one potential-flow body (NBody > 1). To compute the hydro-elastic response of flexible substructures requires the use of HydroDyn coupled within OpenFAST, including the use of the SubDyn model for substructure structural dynamics. That said, SubDyn can output the internal member-level loads (reaction forces and moments within members), but does not compute stress/strain directly. The latter requires post-processing, e.g., using sectional analysis or FEA structural analysis.
Best regards,
Thank you, @Jason.Jonkman
(1) So does the coupled version in OpenFAST allow for a hydro-only analysis? For example, by setting the wind to 0 or even disabling AeroDyn.
(2) In the case of the potential flow solution, how would that look like for the OC4 floater? Would each column a distinct body?
Also, I think I caused some confusion by mentioning the lumped mass approach; In fact, I am interested in the internal member loads of the rigid body semi submersible floater. Is it still necessary then to use SubDyn?
Best,
Luc
Dear @Luc.vanBeek,
Yes, you can always disable AeroDyn (and InflowWind) if you want to simulate the hydro-elastic response of a floating body without aerodynamic loads. You can also disable the tower, drivetrain and blade degrees of freedom of you don’t want a wind turbine atop the floater.
For the OC4-DeepCwind semisubmersible, I would generally suggest modeling each offset column with a distinct potential-flow body and the pontoons, braces, and central column as strip-theory members in HydroDyn. Correspondingly in SubDyn, I would generally suggest modeling each offset column as rigid-link elements and the pontoons, braces, and central column as beam elements. This set-up would allow you to calculate internal member-level loads within the pontoons, braces, and central column. If you also want internal loads within the offset columns, you could further partition each offset column into distinct potential-flow bodies in HydroDyn and model the offset columns as beam elements in SubDyn.
The OC4-DeepCwind semisubmersible is a statically indeterminate structure, so, you can’t calculate internal member-level loads without considering elasticity of the structure.
Best regards,
