Discussion Forum

Hi friends,

I am modeling a composite unit which is made up of Ag and SnO2. I modeling the thermal shock process, as the temperature cooling from 897 K to 297 K and then heating up to 897 K again.

In solid mechanics boundary conditions, I use symmetric condition to one XY, XZ, YZ plane, and use global equation (the intergral of normal stress along the surface is zero) to the other planes to allow them to expand or contract freely but keep flat.

If only the elastic deformation is applied, the model runs well. However, as I add the plasticity to the Ag, the model can fail to converge.

The Convergence Plot 2 is as shown in the attachment, it will decrease to about 1, and then grow to 10e8 and stay, which means it is almost stop and not converge

I find this phenomina appears as the temperature difference exceeds about 200 K, and has nothing to do with the cooling rate. For example, if I model the first 10s of the cooling from 897K to 297K with a thermal transport coefficient K equal to 10. The model will converge, and the T at 10 s is about 680K. However, if I model the 100s (so the sample will cooling to a lower temeprature), the time reciprocal will reach to 10e8, which is not converge. Or if I change the K as 100, model the first 1s, the results converge. But if I model the 10s, the result diverges.

Despite how I change the combiniation of K and time, the Covergence Plot 2 keeps in same shape as decrease to 1, and then boost up to 10e8.

I also tried to change the hardening model, but even at perfect plasticity, this phenomina still exists.

I feel this will always appear as the temperature cooling more than some specific range (such as 200 K).

Does anyone have some idea about this problem? Should I modify the solver selections?

Thanks