Benchmark Davies et al. case 2.3: in.temperature and solver

ASPECT
1

Dear all,

I am modeling a convection flow where I need to implement a viscosity-temperature dependent equation; using as a reference the script of the case 2.3 in -The 2D cylindrical shell benchmarks by Davies et al.- I wrote the equation: out.viscosities[i] = ref_visc * std::min(delta_visc,exp(A*((1.0/in.temperature[i])-1.0))); and created the new library but when I try to run the input with the new add share library I received the error: Floating point exception (core dumped) and it’s due to the (in.temperature) variable. I don’t understand very well why this division is giving me troubles, please, help me with this problem. <ref_visc, delta_visc and A are constants>
Another problem that I am having while I try to run the inputs of the case 2.1 and case 2.3 is that after the timestep 1 in the case 2.1 the solver says that did not converge. In this case I am using the viscosity-temperature dependent equation of the manual model but I can not reproduce any result. I attached the error that I received from the simulation of the case2.3 .
Please, help me with these doubts.

error
error2560×1440 409 KB

Thanks in advance,
Kind regards,
Mark

2

Dear Mark,
Thank you for reaching out on the forum!

I haven’t looked up the cylindrical shell benchmarks, but in the snapshot you attach I see that the minimum temperature is 0. Division by zero could give you the floating point exception.

The outputted velocities seem really high, are they what you expect?

Let me know whether you still need help on this subject!
Cheers,
Anne