How to Use the Optimization Tool

[tibovogel]

Hi everyone,

I am trying to build a 3D composite plates model so as to get the same 10 first normal modes as those I have observed while performing experiments.

My model is built, the problem is the normal modes I have with nastran are too far from my experiments: I would like to optimize them by finding the best young modulus and poisson's ratio to be closer to the value I have observed. I want to tell nastran "find me the best young modulus and poisson's ratio so that I get 120hz for the 1st normal mode" without changing the thickness or dimensions of my plate.

The goal is to use again this model once he will be accurate enough to forecast impact influence on composite structures.

Can you help me understand how to use the "optimization tool" toward this goal?

Regards,
Thibaut.

 

[gantovnik]

Since you did not mention ply orientation angle, I assume that you speak about composite laminate with isotropic layers. If it is simple plate with isotropic layers you can quickly write your own code based on classical lamination theory. Why Nastran? Even Excel now include genetic algorithm.

 

[hezhj2000]

You can definitely use Nastran for this purpose. Use f1, f2, f3, etc. to note natural frequencies from Nastran simulation. You can set (f1-120)^2 as the objective function. The Young's modulus and Poisson's ratios are design variables. Assume the natural frequencies from test data are ft1=120, ft2, ft3, etc. You can also add (f2-ft2)^2<tolerance, etc. as constraints to make sure higher frequencies match the test data.