Time step
 

Hi all--
I am modeling a stiff reaction network in laminar flow in a microchannel. The channel is not straight and has several mixing regions. I am using fluent and an implicit segregated solver. The nominal fluid residence time in the channel is 1-10 seconds, however some of the reaction rates are on the scale of 1e-6 sec.

I am monitoring the concentration of certain species, and as soon as they asymptote out I think I am done, but if I change the timestep to be lower or higher the species composition changes. Should I keep going lower and lower until it does not change?

If I try and solve it on the time scale of the chemistry I will be here forever.

Thanks,
Aaron