Coupled Temp-displacement Convergence problem

[aslfd]

Hello,

I m analyzing the rapid cooling (about 700C°C/s) of steel plate (thickness 1mm) in Abaqus/Standard.

I have same convergence problem like :

***ISOTROPIC HARDENING PLASTICITY DID NOT CONVERGE FOR ELEMENT PLATE-1
***THE SYSTEM MATRIX HAS 6 NEGATIVE EIGENVALUES

I tried
[ul]
[li]- reduce the size of mesh[/li]
[li]- use automatic stabilization (damping factory)[/li]
[li]- dashpots[/li]
[/ul]

but i can't have total convergence. What else can I try to eliminate convergences problems?

 

[corus]

Negative eigenvalues usually means it's not restrained properly. In addition, for rapid cooling use a very small time step and refine the mesh towards the surface where the temperature change is greatest.