Showing results for 
Search instead for 
Do you mean 
Reply

Substructuring in nastran

[ Edited ]

Hi all!
I edited my first question to break it down to a more clear problem.

 

Okay perhaps i should start with something more simple.

 

I have a structure consisting of two structures connected on one edge.

Now i want to stiffen Structure2 with a factor.

First approach:
Vary Youngs modulus of the Material of Structure2 with factor

Second approach:
Vary the stiffness matrix entries

How to do that: Split the Structure2 into a standalone structure

Step 2: Generate mass and stiffness matrix of Structure2 using all DOF

Factorize this stiffness matrix with the factor and add it to the baseline structure using K2GG = factor*KAAX

In some cases this approach is leading to correct results. sometimes it does not unfortunately. Mainly it is the case on the "boundary" grids. (e.g. circumferential structures)

 

Am I missing a basic theory reason behind it?  Are there aspects which lead to differences between the material and the matrix stiffening approach?

 

I have already created a standalone Matrix for Structure2 modified by young's modulus and compared it to a factorized standalone Matrix for Structure2. These matrices are approximately identical.

The errors are appearing when taking the whole structure into consideration. Are there some DOF couplings that "destroy" this opportunity to scale matrices? Or is the Young's modulus also influencing further entries?

 

could you please help me ?

 

Regards

Benny

 

 

 

3 REPLIES

Re: Substructuring in nastran

Does you model solve without any problem withtout the SE approach?

Have you got contraints that stop rigid body motion?

 

Production: NX9.0.3.4, NX10.0.2.6
Development: VB.NET (amateur level !)

Re: Substructuring in nastran

Thanks for your answer.

 

Do you mean if the entire structure solves without any further modification?
Yes, if i want to solve the entire Structure, no issues are occuring.

No i don't have any mechanisms that constrain rigid body modes. I expect some RBM but all other Eigenfrequencies should have to be identical or close to the original values.

 

 

Re: Substructuring in nastran

Okay perhaps i should start with something more simple.

 

I have a structure consisting of two structures connected on one edge.

Now i want to stiffen Structure2 with a factor.

First approach:
Vary Youngs modulus of the Material of Structure2 with factor

Second approach:
Vary the stiffness matrix entries

How to do that: Split the Structure2 into a standalone structure

Step 2: Generate mass and stiffness matrix of Structure2 using all DOF

Factorize this stiffness matrix with the factor and add it to the baseline structure using K2GG = factor*KAAX

In some cases this approach is leading to correct results. sometimes it does not unfortunately. Mainly it is the case on the "boundary" grids. (e.g. circumferential structures)

Am I missing a basic theory reason behind it?  Are there aspects which lead to differences between the material and the matrix stiffening approach?