Hi,
I'm not sure what you're describing as the problem. In the steady state, almost all of the NOx is being formed in the"main" zone, and remains present through the secondary/dilution zone. This seems reasonable as the temperature in the main zone is also the highest. The NOx peak around 100 ms is associated with the H-radical "igniter", which is a pretty artificial model. I wouldn't ascribe any importance to the emissions behavior during this transient.
For clarity, I would also point out that what you are plotting is the time history for a single simulation which at steady state has one residence time (or, if you like, residence times for each of the three zones). The time axis in the plots is not showing what would happen if you varied the residence time.
Regards,
Ray