Learning Forum

[dhjang4522]

Hi,

I  used the SMAT command to export the mass matrix in modal analysis. However, I noticed that the mass matrix only contains values in the X-direction, unlike the stiffness matrix. Could someone explain why this is the case?

 

Also,  displacement values do not diverge when I conduct a quasistatic analysis ignoring mass ( I used Newmark method *gamma = 0.005). However, when I include mass and use the transient analysis with F(t)=Ku(t)+Mu''(t), the values diverge.

 

[dlooman]

It's the *SMAT command, right?  Could you describe the model that you imported the mass matrix from?  Also, the loading, F(t).

[dhjang4522]

Hi Dave,

/AUX2
! Combine .FULL files
COMBINE, FULL
/POST1
*SMAT, MatMS, D, IMPORT, FULL, file.full, MASS
*PRINT, MatMS, Msparse.txt

 

 

I used the code above. It exports the mass matrix through modal analysis of a cantilever beam with one fixed end. The load applied was (10^5 \sin(200t)) at a single node on the end to Z direction for a transient analysis up to 1 second. While there were no issues in ANSYS, the values diverged significantly when I ran it in MATLAB.

[dlooman]

Why is the combine function needed for such a simple model?

As a "sanity check", does a static analysis of acceleration in y and z produce a zero reaction force?

[dhjang4522]

Hi Dave,

I used code that i found on interent and the same matrix is generated without using it. There are reaction Forces in all directions. I think I might have misunderstood initially; it seems that mass exists not only in the X-direction but also specifically at each degree of freedom. So, should I understand it as when a force is applied in a certain direction, acceleration is produced only in that direction, with the formula Fx1 = M1xA1x + M2xA2x + M2x*A3x  and Fy1 = M1yA1y + M2yA2y + M2y*A3y ...?

[dlooman]

Yes, that would be the FX force equilibrium equation if the model is unsupported (no static or damping forces.)