reducing n-dof vibrating system to 1-dof equivalent

[AC1985]

Hi,

I have designed a vibration controller which alters the thrust force
produced by a wind turbine rotor to control the tower top
displacement. This is my dissertation project.


I now need to validate my model with data from an industrial program.


The problem I have is that the indyustrial program models the tower
with 19 masses connected in series with springs. My model only has one
mass connected to one spring.


I am considering taking the 19-dof tower model and analysing the 1st
mode (the one where the top of the tower has maximum displacement). I
think that if I only consider this mode shape, I can reduce the model
into a one-dof system which represents the dynamic characteristics of
the top of the tower.


Could somebody please tell me if my assumptions are correct or
incorrect before I dive into the calculations?


Thanks,
Adam

 

[GregLocock]

Yes, the first mode is often the most important, when dealing with broadband exciation. BUT if your excitation is at blade passing frequency then you are in trouble.


Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[271828]

It sounds like something that we do in buildings. We have a general FE model with zillions of DOF and use static condensation (Guyan Reduction a better term?) to only include the ones we want for earthquake analysis, etc. Seems to me like it would work.

 

[rob768]

Depends on the stiffness and masses you lump. If there is a relatively weak spring between masses you add up, you find significnatly differences between both models. This method works only if the first mode is governed by a shaft with a low stiffness, compare to all others. If there are more soft shafts in the installation, it will influence the effect of the masses added together (in general, you find a lower natural frequency then when the total system is used)