Application of Moments in Wind Turbine Simulation by Beams

Hello everyone,

I dont know if this is the right place for questions like this, but since there should be a lot people involved with numerical modelling of wind turbines, I thought I might give it a try :slight_smile:

I am currently working on a coupling between a fluid solver and a Multibody System program. The fluid solver is called ‘panMARE’, which is a potential method developed at my university. It can be used in order to calculate forces on arbitrary structures in fluids. The Multibody System is called ‘MBDyn’. Ultimately, wind turbine simulations will be carried out using this coupling.

On the MBDyn side, the tower and blades are represented by beam elements. Since this means a transfer from a ‘standard’ 3D structure to beam elements, one has to put thought into how to assemble (fluid solver) and apply (multibody solver) forces and moments. Also, application of the related deformations requires thought. While the assembly of forces and moments seems pretty straight forward, I am having trouble understanding in which way moments and deformations related to the deformed orientations of the nodes should be applied. I will try to further illustrate this below.

Two concepts crossed my mind

a) compute moments with the middle node of the beam elements as reference, then fully apply them

  • on first thought this seems to be the most physically correct way
  • problematic: fully applied moments change the orientation of the nodes on the MBDyn side in 3 DOFs
    → this new orientation should be applied to the body sections that coincide with the nodes
    → imagining the very extreme case of an orientation turned by 90 degrees illustrates that fully applying moments and the related deformations might lead to an unwanted change in section area

b) compute moments with the middle node of the beam elements as reference, only apply the portion of moments that have a torsion-only effect on each beam element

  • this method clearly does not consider all effects but seems to be more stable

I can’t really decide which method should be used or if there is another even better approach.
Can anyone give some insight on this problem or recommend related literature ?

I hope I could get across my problem … if not feel free to ask further questions !

Best Regards,

Tobi