Implementing SSI on MHK RM 1

Wind & Water Structural Analysis
1

Hello,
I am currently working on incorporating soil–structure interaction (SSI) for the MHK RM 1 turbine. In my simulation, I’m using ExtPtfm to provide a stiffness matrix at the mudline in order to account for foundation flexibility. To achieve this, I modified the TOWER FORE-AFT MODE SHAPES and TOWER SIDE-TO-SIDE MODE SHAPES using BModes. Specifically, I use a polynomial fit of the form:
image
This formulation inherently ignores the displacement at the mudline since the constant (𝑎0) and linear (𝑎1) terms are set to zero. However, if I attempt to include these terms by fitting a full polynomial:
y(x) = a₀ + a₁ * x + a₂ * x² + a₃ * x³ + a₄ * x⁴ + a₅ * x⁵ + a₆ * x⁶
and then input entries for TwFAM#Sh(0) and TwFAM#Sh(1) in the ElastoDyn tower input file (and similarly for side-to-side mode shapes), I receive numerical errors during initialization.
My questions are: Is it possible to provide a full polynomial fit including 𝑎 0 ​ and 𝑎 1 in the ElastoDyn tower input file? If it is, could someone advise on the proper method to input these additional coefficients without producing numerical errors? If not, what is the recommended approach within OpenFAST? Currently, my tower mode shape is defined only from 𝑥 2 to 𝑥 6 , which means that any displacement at the mudline is ignored. I would like to see the effect of applying a stiffness matrix at the mudline (via ExtPtfm) on the overall tower behavior.

2

Dear @Tareq.AbuAgolah,

This question has been discussed in several other forum topics, e.g., see: Turbine-soil interaction. Influence on the mode shapes. - #11 by Jason.Jonkman. Basically, the translation and rotation of the tower at its base, interacting with the soil, should come through the platform degrees of freedom (DOFs) in ElastoDyn, rather than through the tower DOFs. The tower mode shapes should be cantilevered (starting with the a_2 term) to the platform, meaning that the mode shapes should be found such that the shape is tangent to the tower at its base. If you are using the NREL provided utility ModeShapePolyFitting.xls to fit the polynomial coefficients, this process is handled automatically.

Best regards,

1 Like