Dear Dr. Jonkman,
As you recommended, I specified the lumped mass/inertia of RNA in ElastoDyn and used SubDyn to model a tower and substructure. I turn off the other modules. Here, I have some questions for you.
First of all, even if I specify the tower properties in the SubDyn input file, ElastoDyn still needs a tower data file and also geometrical data such as tower height. I just put random numbers to make it run but got to know that the results are significantly affected by those inputs. In this case, which numbers I am supposed to put here if the tower is modeled in SubDyn? Are there any dummy numbers?
As regards the output of SubDyn, I set OutAll to TRUE to get the end forces of all the members. In the out file generated, twelve columns(results) for each end of a member are reported. First six columns are the forces and moments for each DOF; but I am not understanding the other six columns. Looking at the units, it is also forces and moments but I am not understanding what those forces and moments indicate. For instance, for joint 1 of member 1, M1J1FKxe, M1J1FKye, M1J1FKze, M1J1MKxe, M1J1MKye, M1J1MKze are the forces and moments I know; but how about the M1J1FMxe, M1J1FMye, M1J1FMze, M1J1MMxe, M1J1MMye, M1J1MMze?
Because I did not specify any time-varying loads such as wind, I expect that the reaction force at the clamped point(seabed) will not show much variation in time but I was wrong. Even the sign of the horizontal force(X) changes. Looking at the previous postingin the forum, it seems this is in inevitable but I wonder what causes these variations. For your reference, I only used ElastoDyn and SubDyn. In the ElastoDyn input file, all the DOFs and mass and inertia are zeroed except for the platform. In the SubDyn, structural properties and geometries of a tower and substructure are defined.
Thank you as always.
As mentioned in the SubDyn documentation, when modeling the full support structure using SubDyn, the platform reference point in ElastoDyn should be located at the yaw bearing; in this case, the tower-bending DOFs in ElastoDyn should be disabled. Have you done this? This should dictate what values of TowerHt, TowerBsHt, PtfmCMzt, and PtfmRefzt you are using. Assuming you’ve set TowerBsHt = 0, the length of the tower in ElastoDyn would be TowerHt - PtfmRefzt, which I assume would be very close to zero (infinitesimal). The tower properties specified in the ElastoDyn tower input file apply to this short (near zero length) section of the tower. To have this short section of tower not effect anything, I would ensure that the tower-bending DOFs are disabled in ElastoDyn and that the mass per unit length specified in the ElastoDyn tower file is very close to zero.
Regarding the member-level outputs, SubDyn calculates and outputs both the static load (Ku) and the inertial load (Mudd), distinguished by “K” and “M”, respectively, in the member-level load output name. These are also discussed a bit in the SubDyn documentation. The static load (K) should be very close to the reaction load; the inertial load (M) is likely less useful.
Regarding your last question, I assume you are referring to the initial start-up transient caused by gravitational loading (in the absence of wind or waves)? The start-up transient can be reduced by specifying good values for the initial conditions. SubDyn does not allow you to specify nonzero initial conditions for the Craig-Bampton modes, but ElastoDyn does allow you to specify nonzero initial conditions for the platform (tower-top, in your case) degrees of freedom. You can find figure out the appropriate values for the initial platform DOFs by first running a simulation with zero initial conditions and finding out where the platform displacements settle out to over time.