Error with MAT9OR

Numair

Hello Team,

 

I got this below error while doing optimization, and I cannot get the model to run successfully.  If I use PARAM,MATCHECK, No, then I would probably get wrong results. what should I do? I changed my design variables bounds, but that still did no do it. 

 

 

A fatal error has occurred during computations:

  *** ERROR # 1551 ***
  for material id = 21 referenced from property id = 21.
  Material coefficients Gij on MAT9 card produce material matrix with negative
  eigenvalues. This violates material stability conditions.
  Eigenvalues of G: 5.811209e+04 -4.221653e+05 1.943692e+04
                    5.542000e+04 2.917000e+04 3.024000e+04

 *** PROGRAM STOPPED: FATAL ERROR(s) ENCOUNTERED

GTT Johan

Hi Numair
So by alternating bounds do you mean that you have design variables for material coefficients Gij?

/Johan

Numair

Johan Dahlberg_20306 said:

Hi Numair
So by alternating bounds do you mean that you have design variables for material coefficients Gij?

/Johan

I have created Design Variables and allowed the 'E' value to vary in a range for the MAT9ORT.

Numair

Numair said:

I have created Design Variables and allowed the 'E' value to vary in a range for the MAT9ORT.

I just found out that increasing one of the 'E' values lower limit will solve the problem, but it was just kind of educated guess.  There should be something in Optistruct to direct you on what to do.

GTT Johan

Great that you sorted it out!
I guess it could be hard to cover all combinations of instability, that's why the generic text in the .out file.
I normally do min/max overchecks before I do optimisation, then one know the "potential" of the problem/design
Anyway, good luck with your optimisation!
/johan

Adriano_Koga

Johan Dahlberg_20306 said:

Great that you sorted it out!
I guess it could be hard to cover all combinations of instability, that's why the generic text in the .out file.
I normally do min/max overchecks before I do optimisation, then one know the "potential" of the problem/design
Anyway, good luck with your optimisation!
/johan

There are a few stability checks needed for the material

https://2021.help.altair.com/2021.2/hwsolvers/os/topics/solvers/os/material_property_check_mat9_r.htm?zoom_highlight=mat9

Numair

If you use (Param,Checkmat, no), then aren't you solving the model with erroneous eigen values? I know that you can stop the error from happening, but are the results correct in this case?

Numair

the values below belong to the material card 21, these are the G values.

5.542000e+04 2.917000e+04 3.024000e+04

I noticed that increasing the lower bounds on of the design variables of the material card could solve the issue.

JuanManuelBH_

Hi Numair!
I had the same error as yours. I changed my lower E value and this kind of sorted out. However, have you got any clue about how do OptiStruct calculate matrix G? I would like to calculate the eigenvalues of that matrix by myself. In that way we could be closer to a proper solution rather than changing values randomly.
Thank you!