open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello all,

I'm trying to model a tracer that has very stiff source terms. The idea is to treat them implicitly and see whether it gives stable and reasonable results.

In other words, I must solve this equation:

dF/dt + u_i*dF/dx_i = S(F)
S(F) = Cte_1*(F -Cte_2)**(3-n)*F

But do not know how to compute S(F). Should I calculate S(F)≈ S(F*), with F* being old values of F, and then pass it as an argument to CVDFTR.f? or should I do something different?

Many thanks,

Sebastian

Hi
As a first answer I would suggest you to look at the water quality example and see how the sources termes are implemented (source_waq.F)

Regards

Hi Sebastian,

I'm not sure how to do it in Telemac with the existing subroutines.
But I know that using implicit method normally means you need to do the sub iterations:

S(Fn) = Cte_1*(Fn -Cte_2)**(3-n)*Fn
dFn+1/dt + u_i*dFn+1/dx_i = S(Fn)

If you use CVDFTR, this means you have to include it in a loop in order to do the subiteration.

But there may be better ways already implemented somewhere in Telemac code.

Regards,Qilong