Dear Michele and Mohit,
The HydroDyn model of OC4-DeepCwind semi-submersible that we’ve released is a hybrid model containing a potential-flow solution and a strip-theory solution for viscous-drag loads and water ballasting of the members. It is the water ballasting (filled fluid) modeled in HydroDyn that (1) impacts the mass, center of mass, and inertia of the platform in the ElastoDyn file and (2) requires the addition of an additional linear stiffness matrix (AddCLin) in HydroDyn. See section 6.6.3 of the draft “HydroDyn User’s Guide and Theory Manual” for an explanation on why the (4,4) and (5,5) entries of AddCLin should be nonzero when modeling the water ballast in HydroDyn. Eliminating these terms will effectively eliminate the contribution the water ballast has on the roll and pitch restoring of the semi-submersible, which I can certainly understand would lead to capsizing of the system (and a fatal simulation error).
Regarding the statement about “H^2”, section 4.3.14 of the draft “HydroDyn User’s Guide and Theory Manual” mentions that “rotational inertia of the fluid in the member is ignored”. By this, we mean the following. The rotational inertias of a solid cylinder about its center are:
I_Axial = 1/2mr^2
I_Transverse = 1/12m(3*r^2 + h^2)
where,
m = mass of cylinder
r = radius of cylinder
h = height of cylinder
However, HydroDyn treats the filled fluid as a point mass (not a thin circular disk) at each cross section. This means that a fluid-filled member is treated as a thin rod instead of a solid cylinder, leaving only:
I_Axial = 0
I_Transverse = 1/12m(h^2)
(Only the h^2 term is kept.)
I hope that helps.
Best regards,