How to deal with large stiffness gradients in BModes

I’m currently exploring FAST and BModes by setting up a model of an existing turbine, which is analyzed before in a different aero-elastic software package. I entered the structural mechanical properties of this turbine into BModes, but I got unexpected results compared to other structured analysis. For example, a big difference between the FA and SS modes. After a couple of checks it turns out that the steep stiffness gradients in the tabular data are causing the trouble.

The stiffness gradients are there because of the flanges at the end of the tower sections. The local stifness of elements in the flange area is approximately 10 times larger than the rest of the elements.
Unfortunately, i cannot post my input file because of confidentiality. Let say that i’ve got a tower with elements of

{list]class 1 that represent the tower shell
class 2 that represent a flange.

I modelled it as follows around each flange and I did put the element nodes exactly at the same locations as the nodes in the tabular data.

node position        properties

n                    class 1
n+1                  class 1
n+2                  class 2
n+3                  class 2
n+4                  class 1
n+5                  class 1

However, this seems to be wrong. I noticed that in the example file flanges are not modelled.

What is the best way to model a structure with such properties for BModes, regarding the theoretical FEM behind it?
Should i lump the structural properties into the large elements, or is there an easier way by shaping the input file?

I’m not an expert in this area, but if the stiffness is 10x the rest of the tower, maybe you should consider it as rigid and model only the less-flexible portion of the tower.

Thanks. Upto now it seems indeed that the best solution is to remove the high stiffness gradients from the model.